Portfolio Optimization using AI Full Syllabus
Module 1: Introduction to Portfolio Optimization
- What is portfolio optimization?
- Traditional vs AI-based portfolio management
- Objectives: Maximize returns, minimize risk, balance constraints
- Overview of Modern Portfolio Theory (MPT), CAPM & Efficient Frontier
Module 2: Financial Data Collection & Preprocessing
- Sources: Yahoo Finance, Alpha Vantage, Quandl, etc.
- Time series data cleaning (missing values, outliers)
- Return types: Daily return, Log return, Cumulative return
- Risk measures: Volatility, Sharpe Ratio, Value at Risk (VaR)
Module 3: Classical Optimization Techniques
- Mean-Variance Optimization (Markowitz Model)
- Risk Parity and Minimum Variance Portfolios
- Black-Litterman model
- Limitations of rule-based approaches
Module 4: Machine Learning for Asset Allocation
- Regression models to forecast asset returns
- Decision Trees and Random Forests for return classification
- K-Means clustering for asset segmentation
- Principal Component Analysis (PCA) for dimensionality reduction
Module 5: Deep Learning in Portfolio Optimization
- RNN/LSTM for price prediction
- Autoencoders for anomaly detection in asset movements
- Transformer models for multi-asset sequences
- Deep Q-Learning for strategy optimization
Module 6: Reinforcement Learning for Portfolio Management
- Basics of Reinforcement Learning (RL)
- RL vs Supervised Learning in finance
- Implementing DQN (Deep Q Networks) for portfolio rebalancing
- Using PPO/A3C for multi-asset portfolio strategies
- OpenAI Gym + FinRL environments for simulation
Module 7: Risk Modelling & Constraints
- Diversification vs concentration
- Maximum drawdown and CVaR (Conditional Value at Risk)
- Constraints: sector weight limits, leverage, liquidity
- Constraint-aware optimization using AI techniques
Module 8: Algorithmic Portfolio Optimization Systems
- Building a portfolio optimization engine
- Feature engineering for optimization signals
- Optimization libraries: CVXPY, PyPortfolioOpt
- Multi-objective optimization using AI (Pareto frontier)
Module 9: Backtesting & Performance Evaluation
- Backtesting framework with Python
- Key metrics: Alpha, Beta, Information Ratio, Turnover
- Benchmark comparison: S&P 500, Nifty 50, MSCI World
- Walk-forward validation & cross-validation in finance
Module 10: Real-Time Portfolio Rebalancing using AI
- Dynamic portfolio rebalancing strategies
- High-frequency trading vs periodic rebalance
- Integrating market news sentiment (NLP-based signals)
- Event-driven trading (earnings, Fed announcements)
Module 11: Tools & Libraries
- Python, Pandas, NumPy, Matplotlib, Seaborn
- Scikit-learn, TensorFlow/Keras, PyTorch
- PyPortfolioOpt, Quantlib, Backtrader
- APIs: Yahoo Finance, Polygon.io, Bloomberg (if available)
