Portfolio Optimization Platform using Machine Learning Models

Financial Models for Portfolio Optimization

To approach the optimzation problem of achieving a higher return while lower risk, we go back to these classic models:
Black-Litterman Model
Markowitz Model
Hierachical risk parity

View Matrix and Uncertainty Matrix Generation

We used three types of models to generate the best view matrix and uncertainty matrix for Black-Litterman Model.

We picked stock price return as the proxy for the two matrices mentioned above.

Expected Return Prediction:

1. Statistical Finance Models: ARIMA, GARCH
2. Machine Learning Models: Linear Regression, Random Forest, XGBoost
3. Neural Network Model: Monte Carlo Dropout

Baseline Result

Short / Long term suggestion

User Interface

Companies preferences Stock preferences Expected Return Acceptable Risk time??

API

Repositories used:

https://github.com/robertmartin8/PyPortfolioOpt https://nbviewer.org/github/Marigold/universal-portfolios/blob/master/modern-portfolio-theory.ipynb

Stock Price EDA

1. closing price trends
2. daily high
3. trading volume
4. moving averages
5. exponential moving averages

Indicator Analysis

1. Moving Averages (SMA, EMA, Usage - Identifies trends and potential reversal points. Commonly used moving averages are 50-day and 200-day SMAs.)
2. Momentum Indicators: (Relative Strength Index (RSI), Stochastic Oscillator)
3. Trend Indicators: Moving Average Convergence Divergence (MACD), Average Directional Index (ADX):
4. Volatility: Bollinger Bands, Average True Range (ATR)
5. Volume: On-Balance Volume (OBV), Volume Moving Average