April 21, 2025

Overview

In this project, I developed a sophisticated portfolio optimization system that leverages time series forecasting and modern portfolio theory to optimize a $1M stock portfolio. The system incorporates ESG (Environmental, Social, Governance) factors, volatility constraints, and diversification requirements to create optimal asset allocations that maximize risk-adjusted returns while maintaining sustainable investment principles.

Key Features

• Holt-Winters Forecasting: Implemented advanced time series analysis using Holt-Winters exponential smoothing to predict future stock prices and market trends with high accuracy.
• Modern Portfolio Theory: Applied Markowitz optimization techniques to construct efficient frontiers and identify optimal asset allocations based on risk-return profiles.
• ESG Integration: Incorporated Environmental, Social, and Governance criteria into the optimization process to ensure sustainable and responsible investment strategies.
• Risk Management: Developed comprehensive risk assessment models including Value at Risk (VaR) calculations and stress testing under various market scenarios.

Technologies Used

• Python: Core programming language with NumPy, Pandas, and SciPy for numerical computations and data analysis and yfinance to get stock data.
• Machine Learning: Scikit-learn for predictive modeling and time series forecasting techniques.
• Optimization: MOSEK and CVXPY in Julia for solving complex quadratic programming problems in portfolio optimization.

Challenges and Limitations

a. Model sensitivity to estimation errors in returns and covariances b. Assumes market conditions remain relatively stable over the investment pe- riod c. Doesn't include real-world constraints like transaction costs, liquidity, or taxes d. No explicit diversification constraints beyond the natural diversification from covariance relationships

Outcome

Portfolio Performance: Our optimization model yielded the following results: • Model Predictions: – Predicted Portfolio Value: $1,690,375.42 – Predicted Profit: $690,375.41 • Actual Performance: – Actual Portfolio Value: $1,389,619.05 – Actual Profit: $389,619.05 – Net Profit (after loan repayment): $101,147.17 – Final Return on Investment: 10.09%

---

This project presents an integrated solution to maximizing a 2-year stock portfolio with a guaranteed minimum profit target. Our approach demonstrates that by combining robust statistical methods with modern portfolio theory, we can con- struct an investment strategy that achieves the desired 10% profit target while managing risk appropriately. The results validate our aggressive investment approach, showing that the actual profit exceeds the minimum target despite market volatility