Code Repository: Predictive Modeling of Stock Market Trends - A 25-Year Machine Learning Analysis of Dhaka Stock Exchange with Web Application Deployment

The project develops and evaluates machine learning models (Random Forest, XGBoost, LightGBM, and soft-voting ensemble) for binary classification of daily stock trends in the Dhaka Stock Exchange (DSE) using 25 years of historical data (2000-2025).

Key Components:

1. Data Analysis Pipeline:
   - Complete data preprocessing and feature engineering
   - Time-series cross-validation implementation
   - Visualization and result interpretation
   - Model training, hyperparameter tuning, and evaluation

2. Production Web Application:
   - Flask backend with REST API for real-time predictions
   - Interactive frontend (HTML/CSS/JavaScript)
   - Model persistence and feature engineering in production
   - Full-stack deployment configuration

Repository Structure:
- `/data_analysis/` - Research code for model development
- `/web_application/backend/` - Flask API and model serving
- `/web_application/frontend/` - User interface
- `/models/` - Serialized trained models
- `/data/` - Sample data and preprocessing scripts
- `/docs/` - Documentation and setup instructions

Requirements: Python 3.8+, Flask, scikit-learn, XGBoost, LightGBM, pandas, numpy