A machine learning project for predicting Microsoft (MSFT) stock price movements using multiple AI techniques. This project is made for the my AI class.
This project implements and compares three different approaches to predict whether Microsoft stock will go UP or DOWN the next day:
- Logistic Regression - Baseline binary classifier
- Random Forest - Ensemble learning method
- LSTM Neural Network - Deep learning for time series
The model uses technical indicators as input features:
- Daily returns (% change)
- RSI (Relative Strength Index)
- Volatility (rolling standard deviation)
- Source: Yahoo Finance (via
yfinanceAPI) - Period: 10 years of historical data
- Ticker: MSFT (Microsoft Corporation)
- Python 3.x
- TensorFlow/Keras (LSTM)
- Scikit-learn (ML models)AI
- Pandas, NumPy (data processing)
- Matplotlib, Seaborn (visualization)
Open and run model.ipynb in Jupyter Notebook to:
- Download stock data
- Engineer technical features
- Train and evaluate all three models
- Compare performance metrics
The notebook includes:
- Model accuracy scores
- Confusion matrices
- Classification reports
- Performance comparisons
Made with ❤️ by mar1shell