Forecast-driven portfolio allocation with Modern Portfolio Theory
A machine learning and data science solution for Guide Me in Finance (GMF) Investments. This project applies ARIMA and LSTM time series models to predict asset price movements, then optimizes asset allocation with Modern Portfolio Theory (MPT). The strategy is validated through rigorous backtesting against historical benchmarks.
GMF Investments required a data-driven approach to improve portfolio returns while controlling downside risk. By integrating statistical and deep learning models into the investment pipeline, the solution enables more accurate forecasts, risk-adjusted allocation, and client-ready portfolio insights.
Hybrid Forecasting: Combines ARIMA and LSTM models for robust market trend predictions
Automated Workflow: Modular architecture supports extension, automation, and reproducibility
Risk-Aware Optimization: Uses MPT and Efficient Frontier analysis for return vs. volatility trade-offs
Backtesting & Metrics: Evaluates portfolio strategies with MAE, RMSE, Sharpe Ratio, and cumulative returns
Historical financial data sourcing and preprocessing with yfinance and custom pipelines
ARIMA/SARIMA parameter tuning using auto_arima; LSTM deep learning model optimization (layers, epochs, batch size)
Feature engineering for returns, volatility, and Bayesian change point detection
Portfolio optimization with PyPortfolioOpt and validation via historical backtests
Delivered a reproducible ML-driven portfolio optimization framework that improved Sharpe Ratio and cumulative returns versus baseline benchmarks, providing GMF with a scalable, client-ready investment tool.
Explore the complete implementation, documentation, and code on GitHub.