/predicting_stock_prices

Stock/cryptocurrency price prediction demo using machine learning library scikit-learn and support vector regression

Primary LanguagePython

Predicting_stock_prices

Stock/cryptocurrency price Prediction demo using support vectors regression and csv historical data extracted from CoinMarketCap.

Overview

Based on the 'Learn Python for Data Science #3' code by youtuber @Sirajology on YouTube. The code uses the scikit-learn machine learning library to train a support vector regression on a stock price dataset from a csv file that can be generated using CoinMarketCapScraper to predict a future price.

Remark

I prefered to make the prediction based on the day to day price difference since that patch of data is less volatile and more stationary. I also threw out the max an min of the patch since they were outlayers that ruined the graph quality.

Dependencies

Install missing dependencies using pip

Demo Usage

Once you have your dependencies installed via pip, run the demo script in terminal via

python demo.py

Screenshots

This demo is tested under ubuntu 17 and 18. Moreover the shown screenshots are the results of a run based on the Bitcoin historical data (Jan 1,2018-May 5,2018).

Credits