Stock Prediction Using Machine Learning Algorithms with Special Reference to Technical Indicators

Abstract

Machine learning algorithms have increasingly become chosen tools for stock price prediction. Using a variety of financial news as an input to compare various algorithms for accuracy level has been extensively studied. However, taking some of the prominent technical indicators as an input to test the algorithms' prediction accuracy for a stock price has remained less explored. This study focuses on using chosen seventeen technical indicators to compare selected algorithms to test the prediction accuracy for six Indian stocks as a sample. This study covers the critical time period of the outbreak of the Covid-19 pandemic and attempts to capture the impact on accuracy levels of algorithms. Three algorithms are tested, and among them random forest algorithm has demonstrated superior results. Based on these results, this study proposes a framework to create a platform for further application.