Quantitative Analysis using Machine Learning

As a Data and Statistics enthusiast, I apply all the theoretical knowledge I've gathered to analyze stock market data and extract meaningful patterns and forecasts. I mostly play with the S&P Index data and individual Stock price data.

So far:

1. I have explored an Unsupervised Learning approach for Algorithmic trading. Using volatility analysis, unsupervised learning can cluster similar stocks for anomaly detection and feature optimization. I have utilized the following techniques to bring new features and technical indicators to the data:
   • Garman-Klass volatility.
   • Relative Strength Index (RSI).
   • Bollinger Bands.
   • MACD (Moving Average Convergence Divergence).
   • Fama-French three-factor model.
2. I have also devised an investment strategy using graph theory, which utilizes Sharpe ratios to diversify stock investments across multiple uncorrelated stocks selected from a set of correlated assets.

The addition of these features has improved the clustering mechanism. For instance, instead of random centering, RSI-based clustering has improved the cluster formations.

For a better understanding, you may refer to these videos 1, Ritvikmath