India Vix Stock Price Prediction

Stock price stock footage

This was one of my first projects. My method was to use Support Vector Regressor and Random Forests to do price predictions. (both well known for handling volatile data)

Esentially, I have converted the dataframe to a time series and observed the price changes in a scatterplot to get a sensing of the price change. There, I was able to see a cyclical pattern and extreme volatility during 2020. (probably due to Covid-19)

scatterplot on change in price
afterwards, random forest and SVR was used. Random Forest method yielded a high accuracy and was the clear winner.

However, there are some improvements to be made:

• Using cross-validation/hyper-param tuning to fine tune the model
• Employing ensemble-based learning to yield better result (i.e. bagging or boosting methods)
• data preprocessing
• using other metrics to compare (e.g. R-squared, MAE, MSE, RMSE, etc.)
  github notebook