Portfolio Optimization:

This project demonstrates portfolio optimization using Weighted Correlation Analysis (WCA) and machine learning to predict stock returns and optimize asset allocation. -The implementation utilizes Python, open-source libraries, and real-time financial data from Yahoo Finance.
Lets see how this work from (Rush 1998).

Dependancy:

Show the code

pip install numpy pandas matplotlib seaborn yfinance scikit-learn cvxpy PyPortfolioOpt

Requirement already satisfied: numpy in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (1.26.4)Note: you may need to restart the kernel to use updated packages.

Requirement already satisfied: pandas in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (2.2.2)
Requirement already satisfied: matplotlib in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (3.10.1)
Requirement already satisfied: seaborn in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (0.13.2)
Requirement already satisfied: yfinance in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (0.2.65)
Requirement already satisfied: scikit-learn in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (1.6.1)
Requirement already satisfied: cvxpy in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (1.7.2)
Requirement already satisfied: PyPortfolioOpt in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (1.5.6)
Requirement already satisfied: python-dateutil>=2.8.2 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from pandas) (2.9.0.post0)
Requirement already satisfied: pytz>=2020.1 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from pandas) (2024.2)
Requirement already satisfied: tzdata>=2022.7 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from pandas) (2024.1)
Requirement already satisfied: contourpy>=1.0.1 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from matplotlib) (1.2.1)
Requirement already satisfied: cycler>=0.10 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from matplotlib) (0.12.1)
Requirement already satisfied: fonttools>=4.22.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from matplotlib) (4.51.0)
Requirement already satisfied: kiwisolver>=1.3.1 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from matplotlib) (1.4.5)
Requirement already satisfied: packaging>=20.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from matplotlib) (23.2)
Requirement already satisfied: pillow>=8 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from matplotlib) (10.3.0)
Requirement already satisfied: pyparsing>=2.3.1 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from matplotlib) (3.1.2)
Requirement already satisfied: requests>=2.31 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (2.32.4)
Requirement already satisfied: multitasking>=0.0.7 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (0.0.11)
Requirement already satisfied: platformdirs>=2.0.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (4.3.3)
Requirement already satisfied: frozendict>=2.3.4 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (2.4.4)
Requirement already satisfied: peewee>=3.16.2 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (3.17.6)
Requirement already satisfied: beautifulsoup4>=4.11.1 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (4.12.3)
Requirement already satisfied: curl_cffi>=0.7 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (0.13.0)
Requirement already satisfied: protobuf>=3.19.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (4.25.3)
Requirement already satisfied: websockets>=13.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from yfinance) (15.0.1)
Requirement already satisfied: scipy>=1.6.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from scikit-learn) (1.13.0)
Requirement already satisfied: joblib>=1.2.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from scikit-learn) (1.4.2)
Requirement already satisfied: threadpoolctl>=3.1.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from scikit-learn) (3.5.0)
Requirement already satisfied: osqp>=0.6.2 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from cvxpy) (1.0.4)
Requirement already satisfied: clarabel>=0.5.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from cvxpy) (0.11.1)
Requirement already satisfied: scs>=3.2.4.post1 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from cvxpy) (3.2.8)
Requirement already satisfied: ecos<3.0.0,>=2.0.14 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from PyPortfolioOpt) (2.0.14)
Requirement already satisfied: plotly<6.0.0,>=5.0.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from PyPortfolioOpt) (5.24.1)
Requirement already satisfied: tenacity>=6.2.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from plotly<6.0.0,>=5.0.0->PyPortfolioOpt) (8.5.0)
Requirement already satisfied: soupsieve>1.2 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from beautifulsoup4>=4.11.1->yfinance) (2.6)
Requirement already satisfied: cffi in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from clarabel>=0.5.0->cvxpy) (1.17.1)
Requirement already satisfied: certifi>=2024.2.2 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from curl_cffi>=0.7->yfinance) (2025.4.26)
Requirement already satisfied: pycparser in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from cffi->clarabel>=0.5.0->cvxpy) (2.22)
Requirement already satisfied: jinja2 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from osqp>=0.6.2->cvxpy) (3.1.3)
Requirement already satisfied: setuptools in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from osqp>=0.6.2->cvxpy) (65.5.0)
Requirement already satisfied: six>=1.5 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from python-dateutil>=2.8.2->pandas) (1.16.0)
Requirement already satisfied: charset_normalizer<4,>=2 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from requests>=2.31->yfinance) (3.3.2)
Requirement already satisfied: idna<4,>=2.5 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from requests>=2.31->yfinance) (3.7)
Requirement already satisfied: urllib3<3,>=1.21.1 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from requests>=2.31->yfinance) (2.2.1)
Requirement already satisfied: MarkupSafe>=2.0 in c:\users\rush\appdata\roaming\python\python310\lib\site-packages (from jinja2->osqp>=0.6.2->cvxpy) (2.1.5)


[notice] A new release of pip is available: 25.1.1 -> 25.2
[notice] To update, run: python.exe -m pip install --upgrade pip

Data Collection

We collected historical adjusted closing prices for six stocks (AAPL, MSFT, GOOGL, AMZN, TSLA, META) using Yahoo Finance: \[ R_t = \frac{P_t - P_{t-1}}{P_{t-1}} \] \[ where, R_t: Daily returns, P_t:Adjusted closing price at time t. \]

Show the code

# Data collection:
import yfinance as yf
import pandas as pd
import numpy as np

tickers = ["AAPL", "MSFT", "GOOGL", "AMZN", "TSLA", "META"]

data = yf.download(tickers, start="2020-11-23", end="2024-11-23")['Close']
# big_data = yf.download("AAPL",  start="2018-11-23", end="2024-11-23")
print(data)
returns = data.pct_change().dropna()
returns_v = data.diff().dropna()
print(returns.head())

C:\Temp\ipykernel_21892\1772283861.py:8: FutureWarning:

YF.download() has changed argument auto_adjust default to True

[                       0%                       ][                       0%                       ][                       0%                       ][**********************67%*******                ]  4 of 6 completed[**********************83%***************        ]  5 of 6 completed[*********************100%***********************]  6 of 6 completed

Ticker            AAPL        AMZN       GOOGL        META        MSFT  \
Date                                                                     
2020-11-23  110.934975  154.919495   85.786316  266.785278  201.986221   
2020-11-24  112.221169  155.903000   87.590881  275.223297  205.591248   
2020-11-25  113.059158  159.253494   87.602295  273.901459  205.600861   
2020-11-27  113.604843  159.766998   88.738953  276.107849  206.908249   
2020-11-30  116.001846  158.401993   87.119133  275.273010  205.793137   
...                ...         ...         ...         ...         ...   
2024-11-18  227.213425  201.699997  174.530319  552.662231  412.661926   
2024-11-19  227.472488  204.610001  177.337936  559.331299  414.676819   
2024-11-20  228.189941  202.880005  175.207321  563.747375  412.393921   
2024-11-21  227.711655  198.380005  166.893997  561.325073  410.613739   
2024-11-22  229.056870  197.119995  164.036591  557.387390  414.721161   

Ticker            TSLA  
Date                    
2020-11-23  173.949997  
2020-11-24  185.126663  
2020-11-25  191.333328  
2020-11-27  195.253326  
2020-11-30  189.199997  
...                ...  
2024-11-18  338.739990  
2024-11-19  346.000000  
2024-11-20  342.029999  
2024-11-21  339.640015  
2024-11-22  352.559998  

[1007 rows x 6 columns]
Ticker          AAPL      AMZN     GOOGL      META      MSFT      TSLA
Date                                                                  
2020-11-24  0.011594  0.006348  0.021036  0.031629  0.017848  0.064252
2020-11-25  0.007467  0.021491  0.000130 -0.004803  0.000047  0.033527
2020-11-27  0.004827  0.003224  0.012975  0.008055  0.006359  0.020488
2020-11-30  0.021099 -0.008544 -0.018254 -0.003024 -0.005389 -0.031002
2020-12-01  0.030827  0.016427  0.023347  0.034588  0.009997  0.030233

Weighted Correlation Analysis (WCA):

The Weighted Correlation Score calculates the relationship between stocks while incorporating predefined weights () for each stock: \[ WCA = {w^T}{C_w} \] w:Weight vector , C:Correlation matrix of stock return.

Show the code

# Define weights based on arbitrary criteria (e.g., market cap or equal weights)
weights = np.array([0.2, 0.2, 0.2, 0.2, 0.1, 0.1])

# Correlation matrix
# correlation_matrix = returns.corr()
correlation_matrix = returns_v.corr()

weighted_correlation_scores = weights @ correlation_matrix @ weights.T
print("Weighted Correlation Score:", weighted_correlation_scores)

Weighted Correlation Score: 0.6065165415946983

Feature Engineering

Added technical indicators like Simple Moving Average (SMA) and volatility for feature enrichment. \[ SMA_{20}(t) = \frac{1}{20} \sum_{i=t-19}^t P_i \] \[ Volatility_{20}(t) = \sqrt{\frac{1}{20} \sum_{i=t-19}^t (P_i - \mu)^2} \]

Show the code

# Feature Engineering
def add_indicators(data):
      indicators = pd.DataFrame(index=data.index)
      for ticker in data.columns:
            indicators[f"{ticker}_SMA_20"] = data[ticker].rolling(window=20).mean()
            indicators[f"{ticker}_Volatility"] = data[ticker].rolling(window=20).std()
      return indicators

# Indicators
indicators = add_indicators(data)
indicators = indicators.dropna()
aligned_returns = returns_v.loc[indicators.index]

print("Indicators shape:", indicators.shape)
print("Returns shape:", aligned_returns.shape)
print(indicators.tail())

Indicators shape: (988, 12)
Returns shape: (988, 6)
            AAPL_SMA_20  AAPL_Volatility  AMZN_SMA_20  AMZN_Volatility  \
Date                                                                     
2024-11-18   226.633754         3.998908   198.556999         9.721216   
2024-11-19   226.269009         3.522264   199.302499         9.576879   
2024-11-20   226.193955         3.460969   200.210999         8.961807   
2024-11-21   226.104443         3.393403   200.810999         8.369185   
2024-11-22   226.040386         3.320591   201.275499         7.852670   

            GOOGL_SMA_20  GOOGL_Volatility  META_SMA_20  META_Volatility  \
Date                                                                       
2024-11-18    171.642056          5.975412   573.423947        11.957740   
2024-11-19    172.288206          5.850177   572.381229        12.242935   
2024-11-20    172.945309          5.358929   572.472443        12.167481   
2024-11-21    173.189731          4.927845   572.238678        12.342028   
2024-11-22    173.164342          4.975805   571.535388        12.782058   

            MSFT_SMA_20  MSFT_Volatility  TSLA_SMA_20  TSLA_Volatility  
Date                                                                    
2024-11-18   418.098119         7.656979   281.000498        40.999921  
2024-11-19   417.615739         7.547193   287.401998        40.634043  
2024-11-20   417.163632         7.577066   293.820998        38.451728  
2024-11-21   416.616064         7.637712   297.778999        38.910571  
2024-11-22   416.104144         7.388091   301.947498        40.133145

Train Machine Learning Models:

Machine Learning for Return Prediction Model: Random Forest Regressor Features (𝑋): SMA, volatility, and other indicators Target, (y): Stock returns Training and Prediction Process: Split data into training (80%) and testing (20%) sets. Train the model: \[ \hat{y} = f(X; \theta) \] \[ MSE = \frac{1}{n} \sum_{i=1}^{n} (y_i - \hat{y}_i)^2 \]

Show the code

# Traning model
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestRegressor
from sklearn.metrics import mean_squared_error

X = indicators.values
# y = returns.shift(-1).dropna().values  # Predict next day's returns
y = aligned_returns.values

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
# Train Random Forest model:
model = RandomForestRegressor()
model.fit(X_train, y_train)

y_pred = model.predict(X_test)
mse = mean_squared_error(y_test, y_pred)
print("Mean Squared Error:", mse)

Mean Squared Error: 33.947697558336245

Portfolio Optimization

Portfolio optimization aims to allocate asset weights w to maximize the Sharpe Ratio. Using predicted returns and a covariance matrix (Σ), optimize weights (w) to maximize the Sharpe Ratio. \[ \text{Sharpe Ratio}= \frac{w^T \mu}{\sqrt{w^T \Sigma w}} \] Where, \[ \mu: \text{Expected returns vector, and derived using mean historical returns } \] \[ \Sigma: \text{Covariance matrix of stock returns, and derived using the covariance of returns} \] \[ w:\text{Portfolio weights (sum to 1)} \]

Optimization is performed using genetic algorithms (GA). -Solve for 𝑤 using PyPortfolioOpt to maximize the Sharpe Ratio -Portfolio_performance measure expected return, volatility, and Sharpe ratio

Show the code

# Portfolio performance
from pypfopt import EfficientFrontier, risk_models, expected_returns

# Calculate expected returns and risk model
mu = expected_returns.mean_historical_return(data)
S = risk_models.sample_cov(data)

# Perform mean-variance optimization
ef = EfficientFrontier(mu, S)
weights = ef.max_sharpe()
cleaned_weights = ef.clean_weights()

# Portfolio performance
performance = ef.portfolio_performance(verbose=True)
print("Optimized Weights:", cleaned_weights)

Expected annual return: 19.8%
Annual volatility: 24.2%
Sharpe Ratio: 0.82
Optimized Weights: OrderedDict([('AAPL', 0.46257), ('AMZN', 0.0), ('GOOGL', 0.0), ('META', 0.0), ('MSFT', 0.53743), ('TSLA', 0.0)])

Backtesting: Evaluating portfolio performance over historical data.

Backtesting evaluates the portfolio’s historical performance: 1. Portfolio Returns: \[ R_{t}^{Portfolio} = \sum_{i=1}^{n} w_i R_{t}^{(i)} \] \[ \text{Where } R_{t}^{(i)}: \text{Returns of assest } i \] 2. Cumulative Returns: \[ Cumulative \ Return = \prod_{t=1}^{T} (1 + R_{t}^{Portfolio}) \]

Show the code

import matplotlib.pyplot as plt
# Backtesting
portfolio_returns = (returns * pd.Series(cleaned_weights)).sum(axis=1)
# Cumulative returns
cumulative_returns = (1 + portfolio_returns).cumprod()

plt.figure(figsize=(10, 6))
plt.plot(cumulative_returns, label="Portfolio")
plt.title("Portfolio Cumulative Returns")
plt.xlabel("Date")
plt.ylabel("Cumulative Return")
plt.legend()
plt.show()

Additional Financial Analysis

Incorporate advanced metrics like Value at Risk (VaR) and Conditional VaR (CVaR). 1. Value at Risk (VaR): Measures potential loss over a given time frame with a confidence level (𝛼): \[ VaR_{\alpha} = Percentile_{(1-\alpha) \cdot 100}(R_{t}^{Portfolio}) \] 2. Conditional Value at Risk (CVaR): Average loss beyond the VaR threshold: \[ CVaR_{\alpha} = E[R_{t}^{Portfolio} | R_{t}^{Portfolio} \leq VaR_{\alpha}] \]

Show the code

#  Financial Analysis
# Value at Risk (VaR)
def calculate_var(returns, confidence_level=0.95):
      return np.percentile(returns, (1 - confidence_level) * 100)
# Conditional Value at Risk (CVaR)
def calculate_cvar(returns, confidence_level=0.95):
      var = calculate_var(returns, confidence_level)
      return returns[returns <= var].mean()

var = calculate_var(portfolio_returns, 0.95)
cvar = calculate_cvar(portfolio_returns, 0.95)
print("Value at Risk (95%):", var)
print("Conditional Value at Risk (95%):", cvar)

Value at Risk (95%): -0.02555226122083683
Conditional Value at Risk (95%): -0.03341942270065758

Automate with Real-Time Data

Update the pipeline for real-time data to to periodically rebalance the portfolio.

Show the code

# Re balancing portfolio:
def rebalance_portfolio():
      new_data = yf.download(tickers, period="1y")['Close']
      updated_returns = new_data.pct_change().dropna()
      # Update weights
      mu = expected_returns.mean_historical_return(new_data)
      S = risk_models.sample_cov(new_data)
      ef = EfficientFrontier(mu, S)
      updated_weights = ef.max_sharpe()
      return ef.clean_weights()

new_weights = rebalance_portfolio()
print("Updated Portfolio Weights:", new_weights)

C:\Temp\ipykernel_21892\532968446.py:3: FutureWarning:

YF.download() has changed argument auto_adjust default to True

[                       0%                       ][****************      33%                       ]  2 of 6 completed[**********************50%                       ]  3 of 6 completed[**********************50%                       ]  3 of 6 completed[**********************83%***************        ]  5 of 6 completed[*********************100%***********************]  6 of 6 completed

Updated Portfolio Weights: OrderedDict([('AAPL', 0.0), ('AMZN', 0.0), ('GOOGL', 0.92565), ('META', 0.0223), ('MSFT', 0.0), ('TSLA', 0.05205)])

Experiment with New Models

Experiment with New Models: LSTM for Time-Series Analysis: LSTMs are used for time-series predictions by learning temporal dependencies in sequential data. This model consist two stacked LSTM layers and one dense layer for output. Input shape is (T,1), where T is time window.

Minimize the Mean Squared Error (MSE) loss function: \[ MSE = \frac{1}{N} \sum_{i=1}^{N} (y_i - \hat{y}_i)^2 \] Where, y: actual value, yi:prediction value, N:number of observations

Tried different ML models (LSTMs for time-series data) to analyze performance improvements. For example:

Show the code

# Train simple LSTM model:
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import LSTM, Dense
# SimpleLSTM model:
model = Sequential([
      LSTM(50, return_sequences=True, input_shape=(X_train.shape[1], 1)),
      LSTM(50),Dense(6)
])
model.compile(optimizer='adam', loss='mse')
model.fit(X_train.reshape(-1, X_train.shape[1], 1), y_train, epochs=25, batch_size=32)

Epoch 1/25

C:\Users\Rush\AppData\Roaming\Python\Python310\lib\site-packages\keras\src\layers\rnn\rnn.py:204: UserWarning:

Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.

 1/25 ━━━━━━━━━━━━━━━━━━━━ 1:13 3s/step - loss: 32.5752

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 32.0584 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 32.7541

25/25 ━━━━━━━━━━━━━━━━━━━━ 3s 5ms/step - loss: 32.8880

Epoch 2/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 36.6047

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.3411 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 29.8421

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.5062

Epoch 3/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 24.3362

12/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.3261 

22/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.4293

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.7637

Epoch 4/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 29.3926

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.2777 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.8552

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 32.2372

Epoch 5/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 32.6723

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 28.9974 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.3318

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.0464

Epoch 6/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 40ms/step - loss: 20.2611

10/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 26.1807 

20/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 29.8361

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.7190

Epoch 7/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 15.6891

 9/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 27.2439 

19/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 30.8367

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 31.4908

Epoch 8/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 26.8339

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 29.7430 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.6825

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.2633

Epoch 9/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 39ms/step - loss: 22.5002

 8/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.2637 

15/25 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.4757

22/25 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.2155

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.1028

Epoch 10/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 1s 58ms/step - loss: 33.6553

 8/25 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.2388 

16/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.1586

24/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.7865

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.8527

Epoch 11/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 34ms/step - loss: 88.6320

 9/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 45.9552 

16/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 39.8937

24/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 37.5494

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 37.2480

Epoch 12/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 38ms/step - loss: 24.9062

 9/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.6903 

17/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.1521

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 34.8448

Epoch 13/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 39.9584

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.5758 

20/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 30.9680

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 31.4601

Epoch 14/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 46.1423

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 33.6296 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 32.8666

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 32.9331

Epoch 15/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 51.0158

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 36.8650 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 35.9032

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 35.5340

Epoch 16/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 36.3317

10/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 36.7969 

20/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 36.1774

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 35.6338

Epoch 17/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 27.2199

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 29.5584 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.1462

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 31.6547

Epoch 18/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 28.4602

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 33.1665 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 32.2935

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 32.5366

Epoch 19/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 28.6406

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 28.8082 

20/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 30.3007

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 31.0445

Epoch 20/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 32ms/step - loss: 20.6091

10/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 27.7098 

17/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 28.4481

24/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.5903

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.8981

Epoch 21/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 37ms/step - loss: 35.3641

 9/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 40.5662 

17/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 38.8579

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 37.5914

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 37.4375

Epoch 22/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 41ms/step - loss: 21.1091

 8/25 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 26.2110 

15/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 28.8994

23/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.4413

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.8124

Epoch 23/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 30ms/step - loss: 23.1255

10/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 34.1019 

20/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 33.7564

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 33.7000

Epoch 24/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 21.9882

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 27.6671 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.1517

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 30.8503

Epoch 25/25


 1/25 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 35.1003

11/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 35.7797 

21/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 35.3332

25/25 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 35.0499

<keras.src.callbacks.history.History at 0x20987d4bcd0>

Show the code

# Test simple LSTM:
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score

X_test = X_test.reshape(-1, X_test.shape[1], 1)  #matches input shape
predictions = model.predict(X_test)

# Evaluate
mse = mean_squared_error(y_test, predictions)
mae = mean_absolute_error(y_test, predictions)
r2 = r2_score(y_test, predictions)
print("Mean Squared Error (MSE):", mse)
print("Mean Absolute Error (MAE):", mae)
print("Mean Absolute Error (MAE):", r2)

# Plot
from sklearn.metrics import accuracy_score, classification_report
import matplotlib.pyplot as plt
colors = ['blue', 'green', 'red']  # Blue for predicted, Green for actual, Red for error
companies = ['Apple', 'Amazon', 'Google', 'Meta', 'Microsoft', 'Tesla']
plt.figure(figsize=(12, 20))
for i in range(6):
      plt.subplot(6, 2, 2*i + 1)
      plt.plot(y_test[:, i], label=f"Actual {companies[i]}", color=colors[1], linestyle='-', linewidth=2)  
      plt.plot(predictions[:, i], label=f"Predicted {companies[i]}", color=colors[0], linestyle='--', linewidth=2)  
      plt.title(f"Actual vs Predicted: {companies[i]}")
      plt.xlabel("Time Steps")
      plt.ylabel("Stock Price / Portfolio Value")
      plt.legend()
      # Error
      plt.subplot(6, 2, 2*i + 2)
      error = y_test[:, i] - predictions[:, i]
      plt.plot(error, label=f"Error {companies[i]}", color=colors[2])  
      plt.title(f"Prediction Error: {companies[i]}")
      plt.xlabel("Time Steps")
      plt.ylabel("Error (Actual - Predicted)")
      plt.legend()
plt.tight_layout()
plt.show()

1/7 ━━━━━━━━━━━━━━━━━━━━ 1s 234ms/step

7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 40ms/step 

7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 40ms/step

Mean Squared Error (MSE): 26.47781285129052

Mean Absolute Error (MAE): 3.363066677852618

Mean Absolute Error (MAE): -0.010818519446154917

Complex LSTM Model:

-Added Features: Dropout layers for regularization and hidden layers for enhanced prediction capability. -Hyperparameters: Tuned units, dropout rate, and batch size.

Show the code

# Complex LSTM:Train
import tensorflow.keras.layers as layers
from tensorflow.keras.models import Sequential
# Hyperparameter Tuning
lstm_units = 64  # Adjust based on dataset complexity
dense_units = 16  # Adjust based on prediction task
dropout_rate = 0.2  # Regularization for overfitting prevention
# Model Architecture
model = Sequential([
      layers.LSTM(lstm_units, return_sequences=True, input_shape=(X_train.shape[1], 1)),
      layers.Dropout(dropout_rate),  # Dropout for regularization
      layers.LSTM(lstm_units),
      layers.Dropout(dropout_rate),  # Dropout after second LSTM
      layers.Dense(dense_units, activation='relu'),  # Hidden layer with ReLU activation
      layers.Dense(6)  # Output layer
])
# Model Compilation (consider alternative optimizers or losses if needed)
model.compile(optimizer='adam', loss='mse')
# Model Training (adjust epochs and batch size based on dataset size)
model.fit(X_train.reshape(-1, X_train.shape[1], 1), y_train, epochs=50, batch_size=64)

Epoch 1/50

C:\Users\Rush\AppData\Roaming\Python\Python310\lib\site-packages\keras\src\layers\rnn\rnn.py:204: UserWarning:

Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.

 1/13 ━━━━━━━━━━━━━━━━━━━━ 38s 3s/step - loss: 35.7369

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.6791

13/13 ━━━━━━━━━━━━━━━━━━━━ 3s 8ms/step - loss: 33.0139

Epoch 2/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 34.5054

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.6753 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.8638

Epoch 3/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 42.7630

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 38.0963

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 35.8087 

Epoch 4/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 34ms/step - loss: 28.2206

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.1825 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.0956

Epoch 5/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 30ms/step - loss: 25.1683

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 30.6513 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.0154

Epoch 6/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 31ms/step - loss: 29.1644

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.9927 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.8208

Epoch 7/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 28.1989

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.9429 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.9264

Epoch 8/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 33ms/step - loss: 29.1950

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 29.5882 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.2419

Epoch 9/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 31.5780

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.4993 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.1895

Epoch 10/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 22.6551

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 27.5879 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 30.2029

Epoch 11/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 40ms/step - loss: 24.3820

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 29.4920

12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 31.0782

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 31.4438

Epoch 12/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 31ms/step - loss: 21.9437

 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 29.9720

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 13ms/step - loss: 31.1644

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 14ms/step - loss: 31.8144

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 14ms/step - loss: 31.9444

Epoch 13/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 55ms/step - loss: 31.8550

 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 14ms/step - loss: 29.2790

10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 13ms/step - loss: 30.1696

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 12ms/step - loss: 31.1070

Epoch 14/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 67ms/step - loss: 31.3447

 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 32.4087

12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 33.7544

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 33.7355

Epoch 15/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - loss: 33.0576

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 38.4413 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 36.3062

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 36.1147

Epoch 16/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 41ms/step - loss: 30.7569

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.4926 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.7412

Epoch 17/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 35.7568

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.6613 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 34.2278

Epoch 18/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 35ms/step - loss: 33.0768

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.4377 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.9427

Epoch 19/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 37ms/step - loss: 28.4831

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 30.7343 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.2305

Epoch 20/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 35ms/step - loss: 35.2962

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.4717 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.3989

Epoch 21/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 31ms/step - loss: 30.3430

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.9153 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.4362

Epoch 22/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 24.8695

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.3183 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.3046

Epoch 23/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 31ms/step - loss: 35.2555

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 36.1824 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.4539

Epoch 24/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 37.1997

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 38.0194 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 36.4490

Epoch 25/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 32.2988

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.9886 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.8202

Epoch 26/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 31ms/step - loss: 30.3928

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.9669 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.4808

Epoch 27/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 31ms/step - loss: 30.4895

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.4266 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.9987

Epoch 28/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 23.2250

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.1532 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.3972

Epoch 29/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 32.6150

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.6829 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.4305

Epoch 30/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 30ms/step - loss: 40.6780

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 35.1167 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.3472

Epoch 31/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 30ms/step - loss: 21.7947

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.2674 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.5053

Epoch 32/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 26.8997

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.3256 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.2120

Epoch 33/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 39.6768

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.2513 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.2087

Epoch 34/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 28.1037

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.9397 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.6597

Epoch 35/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 30.3489

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.4186 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.9102

Epoch 36/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 23.5330

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 28.0638 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.0656

Epoch 37/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 28.6698

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.6823 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.2643

Epoch 38/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 29.3437

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.7576 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.0037

Epoch 39/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 28.7408

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.3572 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.9765

Epoch 40/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 30.8203

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.0758 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.1347

Epoch 41/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 38.3761

 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.9926 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.6974

Epoch 42/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 28.6182

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.6136 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.4739

Epoch 43/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 41.8853

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 36.8063 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 35.6666

Epoch 44/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 32ms/step - loss: 24.6639

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.9160 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.0658

Epoch 45/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 26.9875

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.6741 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.3753

Epoch 46/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 43ms/step - loss: 36.3826

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.3762 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.0276

Epoch 47/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 37ms/step - loss: 33.5007

 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.3227 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.4277

Epoch 48/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 31.9555

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.8802 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.7796

Epoch 49/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 20.7801

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 25.2274 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 28.7586

Epoch 50/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 20.3530

 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.4811 

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.4234

<keras.src.callbacks.history.History at 0x2099026e5c0>

Show the code

# Complex LSTM: Test

# Predict on the test data
X_test_reshaped = X_test.reshape(-1, X_test.shape[1], 1)  # Reshape to match LSTM input
predictions = model.predict(X_test_reshaped)
mse = mean_squared_error(y_test, predictions)
mae = mean_absolute_error(y_test, predictions)
r2 = r2_score(y_test, predictions)
print("Mean Squared Error (MSE):", mse)
print("Mean Absolute Error (MAE):", mae)
print("R-squared (R²):", r2)
# Plots
colors = ['blue', 'green', 'red']  # Blue for predicted, Green for actual, Red for error
companies = ['Apple', 'Amazon', 'Google', 'Meta', 'Microsoft', 'Tesla']
plt.figure(figsize=(12, 15))
for i in range(6):
      plt.subplot(6, 2, 2*i + 1)
      plt.plot(y_test[:, i], label=f"Actual {companies[i]}", color=colors[1], linestyle='-', linewidth=2)  
      plt.plot(predictions[:, i], label=f"Predicted {companies[i]}", color=colors[0], linestyle='--', linewidth=2)  
      plt.title(f"Actual vs Predicted: {companies[i]}")
      plt.xlabel("Time Steps")
      plt.ylabel("Stock Price / Portfolio Value")
      plt.legend()
      # Error
      plt.subplot(6, 2, 2*i + 2)
      error = y_test[:, i] - predictions[:, i]
      plt.plot(error, label=f"Error {companies[i]}", color=colors[2])  
      plt.title(f"Prediction Error: {companies[i]}")
      plt.xlabel("Time Steps")
      plt.ylabel("Error (Actual - Predicted)")
      plt.legend()
plt.tight_layout()
plt.show()

1/7 ━━━━━━━━━━━━━━━━━━━━ 1s 244ms/step

7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 38ms/step 

7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 39ms/step

Mean Squared Error (MSE): 26.363624897824725

Mean Absolute Error (MAE): 3.3603088429192964

R-squared (R²): -0.007386329034622626

Scaled complex LSTM:

Show the code

# scaled LSTM:01
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import LSTM, Dense, Dropout, Bidirectional
from sklearn.preprocessing import MinMaxScaler
from sklearn.metrics import mean_squared_error, mean_absolute_error
import matplotlib.pyplot as plt
# Feature Scaling
scaler = MinMaxScaler(feature_range=(0, 1))
X_train_scaled = scaler.fit_transform(X_train.reshape(-1, X_train.shape[1]))  # Scale X_train
X_test_scaled = scaler.transform(X_test.reshape(-1, X_test.shape[1]))  # Scale X_test
# Reshaping the data for LSTM
X_train_scaled = X_train_scaled.reshape(-1, X_train.shape[1], 1)  # LSTM input shape
X_test_scaled = X_test_scaled.reshape(-1, X_test.shape[1], 1)

# Model with Dropout and Bidirectional LSTM
model = Sequential([
    Bidirectional(LSTM(64, return_sequences=True, input_shape=(X_train_scaled.shape[1], 1))),
    Dropout(0.2),  # Dropout layer to prevent overfitting
    LSTM(64, return_sequences=False),
    Dropout(0.2),
    Dense(6)  # Output layer with 6 units corresponding to 6 companies
])

model.compile(optimizer='adam', loss='mse')
model.fit(X_train_scaled, y_train, epochs=50, batch_size=32, validation_split=0.2, verbose=1)
predictions = model.predict(X_test_scaled)
# Evaluate
mse = mean_squared_error(y_test, predictions)
mae = mean_absolute_error(y_test, predictions)
r2 = r2_score(y_test, predictions)
print("Mean Squared Error (MSE):", mse)
print("Mean Absolute Error (MAE):", mae)
print("Mean Absolute Error (MAE):", r2)
# Plots
colors = ['blue', 'green', 'red']  # Blue for predicted, Green for actual, Red for error
companies = ['Apple', 'Amazon', 'Google', 'Meta', 'Microsoft', 'Tesla']
plt.figure(figsize=(12, 15))
for i in range(6):
      plt.subplot(6, 2, 2*i + 1)
      plt.plot(y_test[:, i], label=f"Actual {companies[i]}", color=colors[1], linestyle='-', linewidth=2)  
      plt.plot(predictions[:, i], label=f"Predicted {companies[i]}", color=colors[0], linestyle='--', linewidth=2)  
      plt.title(f"Actual vs Predicted: {companies[i]}")
      plt.xlabel("Time Steps")
      plt.ylabel("Stock Price / Portfolio Value")
      plt.legend()
      # Error
      plt.subplot(6, 2, 2*i + 2)
      error = y_test[:, i] - predictions[:, i]
      plt.plot(error, label=f"Error {companies[i]}", color=colors[2])  
      plt.title(f"Prediction Error: {companies[i]}")
      plt.xlabel("Time Steps")
      plt.ylabel("Error (Actual - Predicted)")
      plt.legend()
plt.tight_layout()
plt.show()

Epoch 1/50

C:\Users\Rush\AppData\Roaming\Python\Python310\lib\site-packages\keras\src\layers\rnn\rnn.py:204: UserWarning:

Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.

 1/20 ━━━━━━━━━━━━━━━━━━━━ 1:15 4s/step - loss: 38.2255

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.1412 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.9946

20/20 ━━━━━━━━━━━━━━━━━━━━ 5s 31ms/step - loss: 34.0629 - val_loss: 30.9838

Epoch 2/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 30ms/step - loss: 21.3417

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.7305 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.4106

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 31.9964 - val_loss: 30.9599

Epoch 3/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 21.0448

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.0331 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.1095

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.2560 - val_loss: 30.9854

Epoch 4/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 29.4902

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.7852 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.7049

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.8576 - val_loss: 30.9690

Epoch 5/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 65.1961

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 43.0074 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 39.1046

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 37.5177 - val_loss: 30.9861

Epoch 6/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 53.9550

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 40.5919 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 37.8716

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 36.8601 - val_loss: 30.9838

Epoch 7/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 25.8934

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.1983 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.6546

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 35.3637 - val_loss: 30.9751

Epoch 8/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 28.7384

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.2838 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.9926

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.7081 - val_loss: 30.9749

Epoch 9/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 33.6411

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.5112 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.7923

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.8835 - val_loss: 30.9670

Epoch 10/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 25.5554

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.0989 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.0608

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.4526 - val_loss: 30.9459

Epoch 11/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 21.0197

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 25.5991 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.6247

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 31.6122 - val_loss: 30.9498

Epoch 12/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 34.5980

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.3850 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.8153

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.9216 - val_loss: 30.9776

Epoch 13/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 65.2428

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 44.4858 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 40.7346

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 39.5249 - val_loss: 30.9947

Epoch 14/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 38.2888

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 42.9128 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 39.6212

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 38.6693 - val_loss: 30.9782

Epoch 15/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 30.6053

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.1316 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.1116

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 34.1463 - val_loss: 30.9569

Epoch 16/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 31.6724

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.5080 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.2704

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 34.4160 - val_loss: 30.9560

Epoch 17/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 20.3457

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 27.9686 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 31.1233

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 31.7427 - val_loss: 30.9680

Epoch 18/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 23.8714

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.9040 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.4630

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.8543 - val_loss: 31.1390

Epoch 19/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 38.3031

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.9482 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.5446

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.6889 - val_loss: 31.0075

Epoch 20/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 26.6943

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 28.0246 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.8725

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 30.7405 - val_loss: 31.0448

Epoch 21/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 37.0437

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 37.4795 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 37.4931

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 36.9063 - val_loss: 31.0703

Epoch 22/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 35.1135

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 34.5994 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.0602

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 34.9161 - val_loss: 31.0286

Epoch 23/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 35.9357

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 36.8587 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.8628

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 35.5382 - val_loss: 31.0846

Epoch 24/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 36ms/step - loss: 21.2480

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 32.9714 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.3978

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 33.4194 - val_loss: 31.0580

Epoch 25/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 32ms/step - loss: 19.2454

 7/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 27.3078 

12/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 29.6018

18/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 31.8523 

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 32.1954 - val_loss: 31.1091

Epoch 26/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 30.0957

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 36.7990 

13/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 36.6308

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 36.0268

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 35.9404 - val_loss: 31.1506

Epoch 27/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 32ms/step - loss: 30.4610

 7/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 40.6306 

14/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 36.8914

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 35.9037 - val_loss: 31.2203

Epoch 28/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 34.1006

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 37.7245 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 38.5760

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 37.6437 - val_loss: 31.2165

Epoch 29/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 21.4602

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.8544 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.4813

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 35.2822 - val_loss: 31.1919

Epoch 30/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 34ms/step - loss: 35.0807

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.1471 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.3879

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.6576 - val_loss: 31.1983

Epoch 31/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 15.6055

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 25.2233 

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 26.4771

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 27.5557 - val_loss: 31.2909

Epoch 32/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 30.7656

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.1680 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.5054

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.0480 - val_loss: 31.2846

Epoch 33/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 69.7142

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 42.4179 

14/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 39.1873

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 37.5252 - val_loss: 31.1631

Epoch 34/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 30.1115

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 28.9725 

14/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 30.0038

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 31.3509 - val_loss: 31.1935

Epoch 35/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 37ms/step - loss: 41.9460

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 40.4155 

14/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 38.8946

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 37.6954 - val_loss: 31.3298

Epoch 36/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 34ms/step - loss: 37.2165

 6/20 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 37.2744

11/20 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 36.2570

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 35.7054

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 12ms/step - loss: 35.4418 - val_loss: 31.1851

Epoch 37/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 35ms/step - loss: 25.6429

 7/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 27.9870 

13/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 29.4767

18/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 30.4832

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 12ms/step - loss: 31.0244 - val_loss: 31.3111

Epoch 38/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 23.1498

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 35.3392 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 34.8586

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 34.3989 - val_loss: 31.2623

Epoch 39/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 32ms/step - loss: 25.5506

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 28.7573 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 30.3734

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 31.5017 - val_loss: 31.2964

Epoch 40/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 33ms/step - loss: 34.4793

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 30.5218 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 29.8318

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 30.7472 - val_loss: 31.3279

Epoch 41/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 36ms/step - loss: 20.6195

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 27.3934 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 29.2202

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 30.5551 - val_loss: 31.3231

Epoch 42/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 31ms/step - loss: 28.6427

 6/20 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 37.2053

11/20 ━━━━━━━━━━━━━━━━━━━━ 0s 12ms/step - loss: 38.2974

17/20 ━━━━━━━━━━━━━━━━━━━━ 0s 11ms/step - loss: 37.5348

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 13ms/step - loss: 36.9747 - val_loss: 31.2875

Epoch 43/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 33ms/step - loss: 39.6387

 7/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.3892 

14/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 31.4217

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 31.9810 - val_loss: 31.3857

Epoch 44/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 42ms/step - loss: 20.7970

 9/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.3976 

16/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 30.5026

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 31.1041 - val_loss: 31.4536

Epoch 45/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 29.6157

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.9808 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.4741

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 33.7657 - val_loss: 31.3054

Epoch 46/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 31.9203

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 28.9810 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 29.5169

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 30.6322 - val_loss: 31.2292

Epoch 47/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 35.7037

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.0139 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 32.5362

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 32.8453 - val_loss: 31.3740

Epoch 48/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 33ms/step - loss: 25.2059

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 33.5752 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 8ms/step - loss: 33.0666

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 33.5181 - val_loss: 31.3247

Epoch 49/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 24.7941

 8/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.3421 

15/20 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - loss: 29.6724

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 30.5503 - val_loss: 31.4887

Epoch 50/50


 1/20 ━━━━━━━━━━━━━━━━━━━━ 0s 28ms/step - loss: 26.8145

 7/20 ━━━━━━━━━━━━━━━━━━━━ 0s 10ms/step - loss: 36.7492

13/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 35.2785 

19/20 ━━━━━━━━━━━━━━━━━━━━ 0s 9ms/step - loss: 34.8877

20/20 ━━━━━━━━━━━━━━━━━━━━ 0s 12ms/step - loss: 34.8095 - val_loss: 31.3682

WARNING:tensorflow:5 out of the last 15 calls to <function TensorFlowTrainer.make_predict_function.<locals>.one_step_on_data_distributed at 0x00000209939E17E0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.


1/7 ━━━━━━━━━━━━━━━━━━━━ 2s 372ms/step

7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 99ms/step 

7/7 ━━━━━━━━━━━━━━━━━━━━ 1s 99ms/step

Mean Squared Error (MSE): 26.7026782580969

Mean Absolute Error (MAE): 3.399179369883035

Mean Absolute Error (MAE): -0.02548891721545543

Show the code

# scaled LSTM:02# Check the shape of X_train and y_train
print(f"X_train shape before reshaping: {X_train.shape}")
print(f"y_train shape: {y_train.shape}")

# Reshape X_train to be (n_samples, time_steps=1, features=12)
X_train = X_train.reshape(X_train.shape[0], 1, X_train.shape[1])  # Reshape to (n_samples, 1, 12)

# Reshape X_test similarly
X_test = X_test.reshape(X_test.shape[0], 1, X_test.shape[1])

# Normalize the data (optional but recommended)
from sklearn.preprocessing import MinMaxScaler

# Fit scaler on X_train and apply to X_train and X_test
scaler_X = MinMaxScaler(feature_range=(0, 1))
X_train_scaled = scaler_X.fit_transform(X_train.reshape(-1, X_train.shape[2]))  # Reshape for scaling
X_test_scaled = scaler_X.transform(X_test.reshape(-1, X_test.shape[2]))  # Apply scaling to X_test

# Reshape back to 3D for LSTM model
X_train_scaled = X_train_scaled.reshape(X_train.shape[0], X_train.shape[1], X_train.shape[2])  # (n_samples, 1, 12)
X_test_scaled = X_test_scaled.reshape(X_test.shape[0], X_test.shape[1], X_test.shape[2])  # (n_samples, 1, 12)

# You can also scale the output if necessary
scaler_y = MinMaxScaler(feature_range=(0, 1))
y_train_scaled = scaler_y.fit_transform(y_train)  # Scaling y_train (6 companies' prices)
y_test_scaled = scaler_y.transform(y_test)  # Scaling y_test (6 companies' prices)

# Check the reshaped data
print(f"X_train_scaled shape: {X_train_scaled.shape}")
print(f"X_test_scaled shape: {X_test_scaled.shape}")
print(f"y_train_scaled shape: {y_train_scaled.shape}")
print(f"y_test_scaled shape: {y_test_scaled.shape}")

X_train shape before reshaping: (790, 12)
y_train shape: (790, 6)
X_train_scaled shape: (790, 1, 12)
X_test_scaled shape: (198, 1, 12)
y_train_scaled shape: (790, 6)
y_test_scaled shape: (198, 6)

Show the code

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import LSTM, Dense, Dropout

# Define model architecture
model = Sequential([
    LSTM(64, return_sequences=True, input_shape=(X_train_scaled.shape[1], X_train_scaled.shape[2])),
    Dropout(0.2),  # Dropout for regularization
    LSTM(64),
    Dropout(0.2),
    Dense(16, activation='relu'),
    Dense(6)  # Output layer with 6 predictions (one per company)
])

# Compile the model
model.compile(optimizer='adam', loss='mse')

# Fit the model
model.fit(X_train_scaled, y_train_scaled, epochs=50, batch_size=64, validation_data=(X_test_scaled, y_test_scaled))

# Predict on X_test_scaled
predictions = model.predict(X_test_scaled)

# Inverse transform the predictions and actual values to get the original scale
predictions = scaler_y.inverse_transform(predictions)
y_test_actual = scaler_y.inverse_transform(y_test_scaled)

# Evaluate the model's performance
from sklearn.metrics import mean_squared_error, mean_absolute_error

mse = mean_squared_error(y_test_actual, predictions)
mae = mean_absolute_error(y_test_actual, predictions)

print(f"Mean Squared Error (MSE): {mse}")
print(f"Mean Absolute Error (MAE): {mae}")

Epoch 1/50

C:\Users\Rush\AppData\Roaming\Python\Python310\lib\site-packages\keras\src\layers\rnn\rnn.py:204: UserWarning:

Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.

 1/13 ━━━━━━━━━━━━━━━━━━━━ 33s 3s/step - loss: 0.2720

13/13 ━━━━━━━━━━━━━━━━━━━━ 3s 28ms/step - loss: 0.2498 - val_loss: 0.2071

Epoch 2/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.2088

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 0.1858 - val_loss: 0.1120

Epoch 3/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 0.1182

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0885 - val_loss: 0.0357

Epoch 4/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0372

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0379 - val_loss: 0.0246

Epoch 5/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0328

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0286 - val_loss: 0.0206

Epoch 6/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0282

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 0.0242 - val_loss: 0.0166

Epoch 7/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0169

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0197 - val_loss: 0.0149

Epoch 8/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0204

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0189 - val_loss: 0.0138

Epoch 9/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0195

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0166 - val_loss: 0.0124

Epoch 10/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0154

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0155 - val_loss: 0.0116

Epoch 11/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0119

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0147 - val_loss: 0.0107

Epoch 12/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0125

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0136 - val_loss: 0.0102

Epoch 13/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0138

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0136 - val_loss: 0.0097

Epoch 14/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0120

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0116 - val_loss: 0.0093

Epoch 15/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0104

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0115 - val_loss: 0.0091

Epoch 16/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0105

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0113 - val_loss: 0.0088

Epoch 17/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0095

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0113 - val_loss: 0.0086

Epoch 18/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0170

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0121 - val_loss: 0.0086

Epoch 19/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0095

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0110 - val_loss: 0.0084

Epoch 20/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0079

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0100 - val_loss: 0.0084

Epoch 21/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0117

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0106 - val_loss: 0.0084

Epoch 22/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 0.0116

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0107 - val_loss: 0.0082

Epoch 23/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0105

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0105 - val_loss: 0.0083

Epoch 24/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0112

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0106 - val_loss: 0.0081

Epoch 25/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0106

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0106 - val_loss: 0.0082

Epoch 26/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0087

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - loss: 0.0100 - val_loss: 0.0082

Epoch 27/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0086

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0096 - val_loss: 0.0082

Epoch 28/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 29ms/step - loss: 0.0124

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0105 - val_loss: 0.0080

Epoch 29/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0110

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0104 - val_loss: 0.0082

Epoch 30/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0105

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 0.0111 - val_loss: 0.0080

Epoch 31/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0118

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0102 - val_loss: 0.0080

Epoch 32/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0098

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0099 - val_loss: 0.0080

Epoch 33/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 0.0083

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0093 - val_loss: 0.0080

Epoch 34/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0094

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0100 - val_loss: 0.0080

Epoch 35/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0073

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0091 - val_loss: 0.0080

Epoch 36/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0100

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0097 - val_loss: 0.0081

Epoch 37/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0112

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0098 - val_loss: 0.0080

Epoch 38/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 0.0105

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0098 - val_loss: 0.0080

Epoch 39/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0077

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0096 - val_loss: 0.0081

Epoch 40/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0088

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0091 - val_loss: 0.0080

Epoch 41/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 0.0093

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0097 - val_loss: 0.0080

Epoch 42/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0061

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0095 - val_loss: 0.0080

Epoch 43/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0089

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0097 - val_loss: 0.0079

Epoch 44/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0104

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0098 - val_loss: 0.0081

Epoch 45/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0085

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0093 - val_loss: 0.0079

Epoch 46/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0097

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 0.0092 - val_loss: 0.0082

Epoch 47/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 0.0081

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - loss: 0.0086 - val_loss: 0.0080

Epoch 48/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 24ms/step - loss: 0.0104

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0096 - val_loss: 0.0081

Epoch 49/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 0.0070

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0088 - val_loss: 0.0081

Epoch 50/50


 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 0.0085

13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 4ms/step - loss: 0.0091 - val_loss: 0.0080

WARNING:tensorflow:5 out of the last 15 calls to <function TensorFlowTrainer.make_predict_function.<locals>.one_step_on_data_distributed at 0x00000209940237F0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.


1/7 ━━━━━━━━━━━━━━━━━━━━ 1s 218ms/step

7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 39ms/step 

7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 39ms/step

Mean Squared Error (MSE): 27.463631442500695

Mean Absolute Error (MAE): 3.4072607113485236

Show the code

import matplotlib.pyplot as plt

# Plot Actual vs Predicted values for each company
companies = ['Apple', 'Amazon', 'Google', 'Meta', 'Microsoft', 'Tesla']
colors = ['blue', 'green', 'red']

plt.figure(figsize=(12, 8))
for i in range(6):
    plt.subplot(3, 2, i+1)
    plt.plot(y_test_actual[:, i], label=f"Actual {companies[i]}", color='green', linestyle='-', linewidth=2)
    plt.plot(predictions[:, i], label=f"Predicted {companies[i]}", color='blue', linestyle='--', linewidth=2)
    plt.title(f"Actual vs Predicted for {companies[i]}")
    plt.xlabel("Time Steps")
    plt.ylabel("Stock Price")
    plt.legend()

plt.tight_layout()
plt.show()

# Plot Error for each company
plt.figure(figsize=(12, 8))
for i in range(6):
    error = y_test_actual[:, i] - predictions[:, i]
    plt.subplot(3, 2, i+1)
    plt.plot(error, label=f"Error {companies[i]}", color='red')
    plt.title(f"Prediction Error for {companies[i]}")
    plt.xlabel("Time Steps")
    plt.ylabel("Error (Actual - Predicted)")
    plt.legend()

plt.tight_layout()
plt.show()

Evaluate and Enhance Portfolio Strategies

-Stress Testing: Simulate extreme market conditions (e.g., 2008 financial crisis, COVID-19 crash) to see how the portfolio performs under stress.

Show the code

# Stress testing:

def stress_test(data, weights):
    stress_returns = data.pct_change().apply(lambda x: x + np.random.normal(0, 0.1), axis=0)
    portfolio_stress_returns = (stress_returns * pd.Series(weights)).sum(axis=1)
    return portfolio_stress_returns

stress_returns = stress_test(data, new_weights)
print("Portfolio returns under stress:", stress_returns.describe())

Portfolio returns under stress: count    1007.000000
mean       -0.011710
std         0.018986
min        -0.102526
25%        -0.021902
50%        -0.010517
75%        -0.001830
max         0.081572
dtype: float64

Alternative Optimization Methods:

-Risk Parity Portfolio: Balance risks across assets instead of focusing on returns. -Black-Litterman Model: Blend market equilibrium with subjective views.

Risk Parity Portfolio

A Risk Parity portfolio balances risk contributions of all assets rather than focusing solely on return.

Show the code

# Risk Parity Portfolio
import numpy as np
import pandas as pd
from pypfopt import risk_models, expected_returns
from scipy.optimize import minimize

# Function to calculate risk contributions
def calculate_risk_contribution(weights, cov_matrix):
    portfolio_variance = np.dot(weights.T, np.dot(cov_matrix, weights))
    marginal_contribution = np.dot(cov_matrix, weights)
    risk_contributions = marginal_contribution * weights / portfolio_variance
    return risk_contributions

# Objective: Minimize the difference in risk contributions
def risk_parity_objective(weights, cov_matrix):
    risk_contributions = calculate_risk_contribution(weights, cov_matrix)
    return np.sum((risk_contributions - np.mean(risk_contributions))**2)

# Risk parity optimization
def optimize_risk_parity(cov_matrix):
    num_assets = cov_matrix.shape[0]
    init_guess = np.ones(num_assets) / num_assets
    bounds = [(0, 1) for _ in range(num_assets)]
    constraints = {'type': 'eq', 'fun': lambda weights: np.sum(weights) - 1}
    
    result = minimize(risk_parity_objective, init_guess, args=(cov_matrix,), bounds=bounds, constraints=constraints)
    return result.x

# Calculate covariance matrix
cov_matrix = risk_models.sample_cov(data)

# Get risk parity weights
risk_parity_weights = optimize_risk_parity(cov_matrix)
print("Risk Parity Weights:", risk_parity_weights)

Risk Parity Weights: [0.20856559 0.1575289  0.18005532 0.1340182  0.20585634 0.11397565]

Performance check after risk parity approach:

Show the code

# Perf check risk parity

portfolio_returns_rp = (returns * risk_parity_weights).sum(axis=1)
cumulative_rp = (1 + portfolio_returns_rp).cumprod()

plt.plot(cumulative_rp, label="Risk Parity Portfolio")
plt.legend()
plt.title("Cumulative Returns - Risk Parity")
plt.show()

Black-Litterman Model

The Black-Litterman model combines market equilibrium (via market cap weights) with subjective views on assets.

Show the code

# Black-Litterman Model

from pypfopt import BlackLittermanModel

# Market cap weights (assuming proportional to market cap)
market_caps = np.array([2.8, 2.2, 1.9, 1.8, 1.3, 0.8])  # example market caps in trillions
market_cap_weights = market_caps / market_caps.sum()

# Define subjective views
views = {
    "AAPL": 0.02,  # Expected 2% return
    "TSLA": 0.04,  # Expected 4% return
}

# Black-Litterman model
bl = BlackLittermanModel(cov_matrix, pi=market_cap_weights, absolute_views=views)
bl_weights = bl.optimize()

print("Black-Litterman Weights:", bl_weights)

Black-Litterman Weights: OrderedDict([('AAPL', 1.7343839402978416), ('AMZN', 0.579792210471359), ('GOOGL', 0.31972315360625647), ('META', -0.18588576276426483), ('MSFT', -1.0881042412215487), ('TSLA', -0.35990930038964375)])

Performance Check

Show the code

# Perf check Black-Litterman Model

portfolio_returns_bl = (returns * pd.Series(bl_weights)).sum(axis=1)
cumulative_bl = (1 + portfolio_returns_bl).cumprod()

plt.plot(cumulative_bl, label="Black-Litterman Portfolio")
plt.legend()
plt.title("Cumulative Returns - Black-Litterman")
plt.show()

Integrating both methods

Show the code

# integrating both methods:

def calculate_metrics(returns):
    sharpe_ratio = returns.mean() / returns.std()
    max_drawdown = (cumulative_returns / cumulative_returns.cummax() - 1).min()
    return {"Sharpe Ratio": sharpe_ratio, "Max Drawdown": max_drawdown}

print("Risk Parity Metrics:", calculate_metrics(portfolio_returns_rp))
print("Black-Litterman Metrics:", calculate_metrics(portfolio_returns_bl))
# print("ESG Metrics:", calculate_metrics(portfolio_returns_esg))

Risk Parity Metrics: {'Sharpe Ratio': 0.052316006470856095, 'Max Drawdown': -0.3214639070169405}
Black-Litterman Metrics: {'Sharpe Ratio': 0.01567599941116814, 'Max Drawdown': -0.3214639070169405}

Incorporate More Data Sources: Future Work

Alternative Data:

Use sentiment analysis from news, social media, or earnings call transcripts. Incorporate macroeconomic indicators (GDP growth, interest rates). Add fundamental data like P/E ratios, earnings growth, or dividends. ## Data Preprocessing: Use NLP techniques to analyze textual data for sentiment signals. Implement feature selection to prioritize impactful predictors.

Monitor and Improve: Future Work

Performance Monitoring:

Track portfolio metrics (e.g., Sharpe ratio, alpha, drawdown) over time. Create performance benchmarks against indices like S&P 500 or NASDAQ. ## Explainability: Use explainable AI techniques (e.g., SHAP, LIME) to interpret model predictions. Identify which features most influence asset selection and rebalancing decisions.

Advanced Modeling:

Deep Learning:

Train LSTMs, GRUs, or Transformers for better time-series predictions. Incorporate reinforcement learning to optimize rebalancing decisions. ## Optimization: Use genetic algorithms or Bayesian optimization to explore non-linear, multi-objective portfolio strategies.

Show the code

# Bayesian optimization:
from geneticalgorithm import geneticalgorithm as ga

assert list(mu.index) == list(S.columns), "The assets in `mu` and `S` must match!"

def fitness(weights):
    weights = np.array(weights)  # Ensure weights are in array form
    weights = weights / weights.sum()  # Normalize weights to sum to 1
    portfolio_return = np.dot(weights, mu.values)  # Use `.values` to extract NumPy array
    portfolio_risk = np.sqrt(np.dot(weights.T, np.dot(S.values, weights)))  # Use `.values`
    return -portfolio_return / portfolio_risk  # Maximize Sharpe ratio

variable_boundaries = np.array([[0, 1]] * len(mu))  # Each weight must be between 0 and 1
algorithm = ga(function=fitness, dimension=len(mu), variable_type='real', variable_boundaries=variable_boundaries)
algorithm.run()
# Extract the results
optimized_weights = algorithm.output_dict['variable']
optimized_weights = np.array(optimized_weights) / sum(optimized_weights)  # Normalize weights
optimized_weights_series = pd.Series(optimized_weights, index=mu.index)
print("Optimized Weights:\n", optimized_weights_series)
print("Sum of Weights:", optimized_weights_series.sum())

__________________________________________________ 0.3% GA is running...__________________________________________________ 0.7% GA is running...__________________________________________________ 1.0% GA is running...|_________________________________________________ 1.3% GA is running...|_________________________________________________ 1.7% GA is running...|_________________________________________________ 2.0% GA is running...|_________________________________________________ 2.3% GA is running...|_________________________________________________ 2.7% GA is running...||________________________________________________ 3.0% GA is running...||________________________________________________ 3.3% GA is running...||________________________________________________ 3.7% GA is running...||________________________________________________ 4.0% GA is running...||________________________________________________ 4.3% GA is running...||________________________________________________ 4.7% GA is running...||________________________________________________ 5.0% GA is running...|||_______________________________________________ 5.3% GA is running...|||_______________________________________________ 5.7% GA is running...|||_______________________________________________ 6.0% GA is running...|||_______________________________________________ 6.3% GA is running...|||_______________________________________________ 6.7% GA is running...||||______________________________________________ 7.0% GA is running...||||______________________________________________ 7.3% GA is running...||||______________________________________________ 7.7% GA is running...||||______________________________________________ 8.0% GA is running...||||______________________________________________ 8.3% GA is running...||||______________________________________________ 8.7% GA is running...||||______________________________________________ 9.0% GA is running...|||||_____________________________________________ 9.3% GA is running...|||||_____________________________________________ 9.7% GA is running...|||||_____________________________________________ 10.0% GA is running...|||||_____________________________________________ 10.3% GA is running...|||||_____________________________________________ 10.7% GA is running...||||||____________________________________________ 11.0% GA is running...||||||____________________________________________ 11.3% GA is running...||||||____________________________________________ 11.7% GA is running...||||||____________________________________________ 12.0% GA is running...||||||____________________________________________ 12.3% GA is running...||||||____________________________________________ 12.7% GA is running...||||||____________________________________________ 13.0% GA is running...|||||||___________________________________________ 13.3% GA is running...|||||||___________________________________________ 13.7% GA is running...|||||||___________________________________________ 14.0% GA is running...|||||||___________________________________________ 14.3% GA is running...|||||||___________________________________________ 14.7% GA is running...||||||||__________________________________________ 15.0% GA is running...||||||||__________________________________________ 15.3% GA is running...||||||||__________________________________________ 15.7% GA is running...||||||||__________________________________________ 16.0% GA is running...||||||||__________________________________________ 16.3% GA is running...||||||||__________________________________________ 16.7% GA is running...||||||||__________________________________________ 17.0% GA is running...|||||||||_________________________________________ 17.3% GA is running...|||||||||_________________________________________ 17.7% GA is running...|||||||||_________________________________________ 18.0% GA is running...|||||||||_________________________________________ 18.3% GA is running...|||||||||_________________________________________ 18.7% GA is running...||||||||||________________________________________ 19.0% GA is running...||||||||||________________________________________ 19.3% GA is running...||||||||||________________________________________ 19.7% GA is running...||||||||||________________________________________ 20.0% GA is running...||||||||||________________________________________ 20.3% GA is running...||||||||||________________________________________ 20.7% GA is running...||||||||||________________________________________ 21.0% GA is running...|||||||||||_______________________________________ 21.3% GA is running...|||||||||||_______________________________________ 21.7% GA is running...|||||||||||_______________________________________ 22.0% GA is running...|||||||||||_______________________________________ 22.3% GA is running...|||||||||||_______________________________________ 22.7% GA is running...||||||||||||______________________________________ 23.0% GA is running...||||||||||||______________________________________ 23.3% GA is running...||||||||||||______________________________________ 23.7% GA is running...||||||||||||______________________________________ 24.0% GA is running...||||||||||||______________________________________ 24.3% GA is running...||||||||||||______________________________________ 24.7% GA is running...||||||||||||______________________________________ 25.0% GA is running...|||||||||||||_____________________________________ 25.3% GA is running...|||||||||||||_____________________________________ 25.7% GA is running...|||||||||||||_____________________________________ 26.0% GA is running...|||||||||||||_____________________________________ 26.3% GA is running...|||||||||||||_____________________________________ 26.7% GA is running...||||||||||||||____________________________________ 27.0% GA is running...||||||||||||||____________________________________ 27.3% GA is running...||||||||||||||____________________________________ 27.7% GA is running...||||||||||||||____________________________________ 28.0% GA is running...||||||||||||||____________________________________ 28.3% GA is running...||||||||||||||____________________________________ 28.7% GA is running...||||||||||||||____________________________________ 29.0% GA is running...|||||||||||||||___________________________________ 29.3% GA is running...|||||||||||||||___________________________________ 29.7% GA is running...|||||||||||||||___________________________________ 30.0% GA is running...|||||||||||||||___________________________________ 30.3% GA is running...|||||||||||||||___________________________________ 30.7% GA is running...||||||||||||||||__________________________________ 31.0% GA is running...||||||||||||||||__________________________________ 31.3% GA is running...||||||||||||||||__________________________________ 31.7% GA is running...||||||||||||||||__________________________________ 32.0% GA is running...||||||||||||||||__________________________________ 32.3% GA is running...||||||||||||||||__________________________________ 32.7% GA is running...||||||||||||||||__________________________________ 33.0% GA is running...|||||||||||||||||_________________________________ 33.3% GA is running...|||||||||||||||||_________________________________ 33.7% GA is running...|||||||||||||||||_________________________________ 34.0% GA is running...|||||||||||||||||_________________________________ 34.3% GA is running...|||||||||||||||||_________________________________ 34.7% GA is running...||||||||||||||||||________________________________ 35.0% GA is running...||||||||||||||||||________________________________ 35.3% GA is running...||||||||||||||||||________________________________ 35.7% GA is running...||||||||||||||||||________________________________ 36.0% GA is running...||||||||||||||||||________________________________ 36.3% GA is running...||||||||||||||||||________________________________ 36.7% GA is running...||||||||||||||||||________________________________ 37.0% GA is running...|||||||||||||||||||_______________________________ 37.3% GA is running...|||||||||||||||||||_______________________________ 37.7% GA is running...|||||||||||||||||||_______________________________ 38.0% GA is running...|||||||||||||||||||_______________________________ 38.3% GA is running...|||||||||||||||||||_______________________________ 38.7% GA is running...||||||||||||||||||||______________________________ 39.0% GA is running...||||||||||||||||||||______________________________ 39.3% GA is running...||||||||||||||||||||______________________________ 39.7% GA is running...||||||||||||||||||||______________________________ 40.0% GA is running...||||||||||||||||||||______________________________ 40.3% GA is running...||||||||||||||||||||______________________________ 40.7% GA is running...||||||||||||||||||||______________________________ 41.0% GA is running...|||||||||||||||||||||_____________________________ 41.3% GA is running...|||||||||||||||||||||_____________________________ 41.7% GA is running...|||||||||||||||||||||_____________________________ 42.0% GA is running...|||||||||||||||||||||_____________________________ 42.3% GA is running...|||||||||||||||||||||_____________________________ 42.7% GA is running...||||||||||||||||||||||____________________________ 43.0% GA is running...||||||||||||||||||||||____________________________ 43.3% GA is running...||||||||||||||||||||||____________________________ 43.7% GA is running...||||||||||||||||||||||____________________________ 44.0% GA is running...||||||||||||||||||||||____________________________ 44.3% GA is running...||||||||||||||||||||||____________________________ 44.7% GA is running...||||||||||||||||||||||____________________________ 45.0% GA is running...|||||||||||||||||||||||___________________________ 45.3% GA is running...|||||||||||||||||||||||___________________________ 45.7% GA is running...|||||||||||||||||||||||___________________________ 46.0% GA is running...|||||||||||||||||||||||___________________________ 46.3% GA is running...|||||||||||||||||||||||___________________________ 46.7% GA is running...||||||||||||||||||||||||__________________________ 47.0% GA is running...||||||||||||||||||||||||__________________________ 47.3% GA is running...||||||||||||||||||||||||__________________________ 47.7% GA is running...||||||||||||||||||||||||__________________________ 48.0% GA is running...||||||||||||||||||||||||__________________________ 48.3% GA is running...||||||||||||||||||||||||__________________________ 48.7% GA is running...||||||||||||||||||||||||__________________________ 49.0% GA is running...|||||||||||||||||||||||||_________________________ 49.3% GA is running...|||||||||||||||||||||||||_________________________ 49.7% GA is running...|||||||||||||||||||||||||_________________________ 50.0% GA is running...|||||||||||||||||||||||||_________________________ 50.3% GA is running...|||||||||||||||||||||||||_________________________ 50.7% GA is running...||||||||||||||||||||||||||________________________ 51.0% GA is running...||||||||||||||||||||||||||________________________ 51.3% GA is running...||||||||||||||||||||||||||________________________ 51.7% GA is running...||||||||||||||||||||||||||________________________ 52.0% GA is running...||||||||||||||||||||||||||________________________ 52.3% GA is running...||||||||||||||||||||||||||________________________ 52.7% GA is running...||||||||||||||||||||||||||________________________ 53.0% GA is running...|||||||||||||||||||||||||||_______________________ 53.3% GA is running...|||||||||||||||||||||||||||_______________________ 53.7% GA is running...|||||||||||||||||||||||||||_______________________ 54.0% GA is running...|||||||||||||||||||||||||||_______________________ 54.3% GA is running...|||||||||||||||||||||||||||_______________________ 54.7% GA is running...||||||||||||||||||||||||||||______________________ 55.0% GA is running...||||||||||||||||||||||||||||______________________ 55.3% GA is running...||||||||||||||||||||||||||||______________________ 55.7% GA is running...||||||||||||||||||||||||||||______________________ 56.0% GA is running...||||||||||||||||||||||||||||______________________ 56.3% GA is running...||||||||||||||||||||||||||||______________________ 56.7% GA is running...||||||||||||||||||||||||||||______________________ 57.0% GA is running...|||||||||||||||||||||||||||||_____________________ 57.3% GA is running...|||||||||||||||||||||||||||||_____________________ 57.7% GA is running...|||||||||||||||||||||||||||||_____________________ 58.0% GA is running...|||||||||||||||||||||||||||||_____________________ 58.3% GA is running...|||||||||||||||||||||||||||||_____________________ 58.7% GA is running...||||||||||||||||||||||||||||||____________________ 59.0% GA is running...||||||||||||||||||||||||||||||____________________ 59.3% GA is running...||||||||||||||||||||||||||||||____________________ 59.7% GA is running...||||||||||||||||||||||||||||||____________________ 60.0% GA is running...||||||||||||||||||||||||||||||____________________ 60.3% GA is running...||||||||||||||||||||||||||||||____________________ 60.7% GA is running...||||||||||||||||||||||||||||||____________________ 61.0% GA is running...|||||||||||||||||||||||||||||||___________________ 61.3% GA is running...|||||||||||||||||||||||||||||||___________________ 61.7% GA is running...|||||||||||||||||||||||||||||||___________________ 62.0% GA is running...|||||||||||||||||||||||||||||||___________________ 62.3% GA is running...|||||||||||||||||||||||||||||||___________________ 62.7% GA is running...||||||||||||||||||||||||||||||||__________________ 63.0% GA is running...||||||||||||||||||||||||||||||||__________________ 63.3% GA is running...||||||||||||||||||||||||||||||||__________________ 63.7% GA is running...||||||||||||||||||||||||||||||||__________________ 64.0% GA is running...||||||||||||||||||||||||||||||||__________________ 64.3% GA is running...||||||||||||||||||||||||||||||||__________________ 64.7% GA is running...||||||||||||||||||||||||||||||||__________________ 65.0% GA is running...|||||||||||||||||||||||||||||||||_________________ 65.3% GA is running...|||||||||||||||||||||||||||||||||_________________ 65.7% GA is running...|||||||||||||||||||||||||||||||||_________________ 66.0% GA is running...|||||||||||||||||||||||||||||||||_________________ 66.3% GA is running...|||||||||||||||||||||||||||||||||_________________ 66.7% GA is running...||||||||||||||||||||||||||||||||||________________ 67.0% GA is running...||||||||||||||||||||||||||||||||||________________ 67.3% GA is running...||||||||||||||||||||||||||||||||||________________ 67.7% GA is running...||||||||||||||||||||||||||||||||||________________ 68.0% GA is running...||||||||||||||||||||||||||||||||||________________ 68.3% GA is running...||||||||||||||||||||||||||||||||||________________ 68.7% GA is running...||||||||||||||||||||||||||||||||||________________ 69.0% GA is running...|||||||||||||||||||||||||||||||||||_______________ 69.3% GA is running...|||||||||||||||||||||||||||||||||||_______________ 69.7% GA is running...|||||||||||||||||||||||||||||||||||_______________ 70.0% GA is running...|||||||||||||||||||||||||||||||||||_______________ 70.3% GA is running...|||||||||||||||||||||||||||||||||||_______________ 70.7% GA is running...||||||||||||||||||||||||||||||||||||______________ 71.0% GA is running...||||||||||||||||||||||||||||||||||||______________ 71.3% GA is running...||||||||||||||||||||||||||||||||||||______________ 71.7% GA is running...||||||||||||||||||||||||||||||||||||______________ 72.0% GA is running...||||||||||||||||||||||||||||||||||||______________ 72.3% GA is running...||||||||||||||||||||||||||||||||||||______________ 72.7% GA is running...||||||||||||||||||||||||||||||||||||______________ 73.0% GA is running...|||||||||||||||||||||||||||||||||||||_____________ 73.3% GA is running...|||||||||||||||||||||||||||||||||||||_____________ 73.7% GA is running...|||||||||||||||||||||||||||||||||||||_____________ 74.0% GA is running...|||||||||||||||||||||||||||||||||||||_____________ 74.3% GA is running...|||||||||||||||||||||||||||||||||||||_____________ 74.7% GA is running...||||||||||||||||||||||||||||||||||||||____________ 75.0% GA is running...||||||||||||||||||||||||||||||||||||||____________ 75.3% GA is running...||||||||||||||||||||||||||||||||||||||____________ 75.7% GA is running...||||||||||||||||||||||||||||||||||||||____________ 76.0% GA is running...||||||||||||||||||||||||||||||||||||||____________ 76.3% GA is running...||||||||||||||||||||||||||||||||||||||____________ 76.7% GA is running...||||||||||||||||||||||||||||||||||||||____________ 77.0% GA is running...|||||||||||||||||||||||||||||||||||||||___________ 77.3% GA is running...|||||||||||||||||||||||||||||||||||||||___________ 77.7% GA is running...|||||||||||||||||||||||||||||||||||||||___________ 78.0% GA is running...|||||||||||||||||||||||||||||||||||||||___________ 78.3% GA is running...|||||||||||||||||||||||||||||||||||||||___________ 78.7% GA is running...||||||||||||||||||||||||||||||||||||||||__________ 79.0% GA is running...||||||||||||||||||||||||||||||||||||||||__________ 79.3% GA is running...||||||||||||||||||||||||||||||||||||||||__________ 79.7% GA is running...||||||||||||||||||||||||||||||||||||||||__________ 80.0% GA is running...||||||||||||||||||||||||||||||||||||||||__________ 80.3% GA is running...||||||||||||||||||||||||||||||||||||||||__________ 80.7% GA is running...||||||||||||||||||||||||||||||||||||||||__________ 81.0% GA is running...|||||||||||||||||||||||||||||||||||||||||_________ 81.3% GA is running...|||||||||||||||||||||||||||||||||||||||||_________ 81.7% GA is running...|||||||||||||||||||||||||||||||||||||||||_________ 82.0% GA is running...|||||||||||||||||||||||||||||||||||||||||_________ 82.3% GA is running...|||||||||||||||||||||||||||||||||||||||||_________ 82.7% GA is running...||||||||||||||||||||||||||||||||||||||||||________ 83.0% GA is running...||||||||||||||||||||||||||||||||||||||||||________ 83.3% GA is running...||||||||||||||||||||||||||||||||||||||||||________ 83.7% GA is running...||||||||||||||||||||||||||||||||||||||||||________ 84.0% GA is running...||||||||||||||||||||||||||||||||||||||||||________ 84.3% GA is running...||||||||||||||||||||||||||||||||||||||||||________ 84.7% GA is running...||||||||||||||||||||||||||||||||||||||||||________ 85.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||_______ 85.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||_______ 85.7% GA is running...|||||||||||||||||||||||||||||||||||||||||||_______ 86.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||_______ 86.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||_______ 86.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||______ 87.0% GA is running...||||||||||||||||||||||||||||||||||||||||||||______ 87.3% GA is running...||||||||||||||||||||||||||||||||||||||||||||______ 87.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||______ 88.0% GA is running...||||||||||||||||||||||||||||||||||||||||||||______ 88.3% GA is running...||||||||||||||||||||||||||||||||||||||||||||______ 88.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||______ 89.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||||_____ 89.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||||_____ 89.7% GA is running...|||||||||||||||||||||||||||||||||||||||||||||_____ 90.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||||_____ 90.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||||_____ 90.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||||____ 91.0% GA is running...||||||||||||||||||||||||||||||||||||||||||||||____ 91.3% GA is running...||||||||||||||||||||||||||||||||||||||||||||||____ 91.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||||____ 92.0% GA is running...||||||||||||||||||||||||||||||||||||||||||||||____ 92.3% GA is running...||||||||||||||||||||||||||||||||||||||||||||||____ 92.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||||____ 93.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||___ 93.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||___ 93.7% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||___ 94.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||___ 94.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||___ 94.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||||||__ 95.0% GA is running...||||||||||||||||||||||||||||||||||||||||||||||||__ 95.3% GA is running...||||||||||||||||||||||||||||||||||||||||||||||||__ 95.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||||||__ 96.0% GA is running...||||||||||||||||||||||||||||||||||||||||||||||||__ 96.3% GA is running...||||||||||||||||||||||||||||||||||||||||||||||||__ 96.7% GA is running...||||||||||||||||||||||||||||||||||||||||||||||||__ 97.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||_ 97.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||_ 97.7% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||_ 98.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||_ 98.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||_ 98.7% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||| 99.0% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||| 99.3% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||| 99.7% GA is running...|||||||||||||||||||||||||||||||||||||||||||||||||| 100.0% GA is running...                                                                                                     The best solution found:
 [8.21011350e-01 2.92045694e-04 3.13948978e-02 1.09605638e-03
 9.55337613e-01 1.05178075e-02]

 Objective function:
 -0.8186701051775898

Optimized Weights:
 Ticker
AAPL     0.451192
AMZN     0.000160
GOOGL    0.017253
META     0.000602
MSFT     0.525012
TSLA     0.005780
dtype: float64
Sum of Weights: 1.0

Sum of weights (0.999999 ~ 1)denotes the total amount of money invested in the portfolio i.e. $1M. Portfolio optimization is not about maximizing the return, but about balancing the risk and return.

Conclusion

Used $1M virtual money to built a portfolio optimization model.This study demonstrated the integration of Weighted Correlation Analysis (WCA) into a portfolio optimization framework. The results indicate that WCA enhances the predictive quality of financial features and can guide more effective allocation strategies compared to baseline methods. The methodology underscores the importance of combining statistical correlation structures with machine learning–based forecasting to improve portfolio outcomes.

Future work

Future research will focus on extending the asset universe beyond equities, incorporating advanced neural forecasting models (e.g., LSTMs, transformers), and introducing risk-aware backtesting frameworks that account for transaction costs and market frictions. Further, the exploration of reinforcement learning and Bayesian models such as Black–Litterman will provide insights into more robust allocation under uncertainty. Ultimately, this line of work aims to bridge the gap between academic innovation and real-world, production-ready investment systems. Compare different weight optimizing methods

References

Rush, Desai R. 1998. “The Quant Book.”