Unsupervised Learning Trading Strategy

• Download/Load SP500 stocks prices data.
• Calculate different features and indicators on each stock.
• Aggregate on monthly level and filter top 150 most liquid stocks.
• Calculate Monthly Returns for different time-horizons.
• Download Fama-French Factors and Calculate Rolling Factor Betas.
• For each month fit a K-Means Clustering Algorithm to group similar assets based on their features.
• For each month select assets based on the cluster and form a portfolio based on Efficient Frontier max sharpe ratio optimization.
• Visualize Portfolio returns and compare to SP500 returns.

All Packages Needed:

pandas, numpy, matplotlib, statsmodels, pandas_datareader, datetime, yfinance, sklearn, PyPortfolioOpt