The project parsed through a data file containing 70 fundamental / momentum attributes (Returns, Volatility, RSI etc) for ~2200 US Equity Tickers and use KNN algorithm to find the nearest neighbor for a given stock and determines its Predicted fundamental ratios (P/E, P/B and P/S). Also generated a 95% confidence range for price using yearly volatility of the equity. Later improvements, as a part of the Data Science capstone project @ General Assembly, took advantage of the Pandas and Gradient Boosted Regression Techniques from Scikit learn.
tastypotinc / pepredictor Goto Github PK
View Code? Open in Web Editor NEWThis project forked from coolcalbeans/pepredictor
The project parsed through a data file containing 70 fundamental / momentum attributes (Returns, Volatility, RSI etc) for ~2200 US Equity Tickers and use KNN algorithm to find the nearest neighbor for a given stock and determines its Predicted fundamental ratios (P/E, P/B and P/S). Also generated a 95% confidence range for price using yearly volatility of the equity. Later improvements, as a part of the Data Science capstone project @ General Assembly, took advantage of the Pandas and Gradient Boosted Regression Techniques from Scikit learn.