Comparative Analysis of Machine Learning Algorithm to Forecast Indian Stock Market

Abstract

Complexity and diversity of the stock market has always attracted the researchers since ages to find out the ways to predict the future movement. However, the volatility across the globe in different stock markets makes this task more difficult. Statistical methods and modelling are effective but can’t address the vast range of problems encountered in the prediction of movement of the stock market. The traditional methods used were not able to address and give solution to the complex problems prevailing in the stock market. The machine learning and artificial intelligence tools are used to solve the complex situations and problems of Big Data with ease. In this paper authors propose to use six different algorithms i.e., Generalized Linear Model, Deep Learning, Decision Tree, Random Forest, Gradient Boosted Trees and Support Vector Machine and identify model which predict near to actuals. These algorithms were applied on the BSE index data from April 2015 to 31st March 2020 and the model with least relative error is identified. Amongst all the models applied, Gradient Boosted Trees is chosen to be the efficient one, as it has least relative error and standard deviation. Further Gradient Boosted Trees is used to forecast the results.