Comparative Study of Machine Learning Models Implemented on Stock Market Datasets

Abstract

Home Proceedings of International Conference on Communication and Computational Technologies Conference paper Comparative Study of Machine Learning Models Implemented on Stock Market Datasets Save Related Papers Chat with paper Priyank Gupta, Sakshi Talreja, Rakesh Singh Jadon & Sanjay Kumar Gupta Conference paper First Online: 27 September 2022 251 Accesses Part of the Algorithms for Intelligent Systems book series (AIS) Abstract Aim of this paper is to analyze some machine learning algorithms,: k-nearest neighbor (KNN), linear regression, random forest (RF), support vector regression (SVR), and implement them on two different datasets of stock market IT sector companies TCS and Wipro. Result shows that the accuracy of the model doesn’t always increase on the size of data. K-fold cross validation is applied for evaluating the performance of the models. Comparison among different ML algorithms is based on training and testing accuracy, training and testing RMSE, and K-fold accuracy using datasets. Results obtained from different algorithms based on R2 score of each dataset are compared using graphs. It suggests that random forest is a best-fitted algorithm among all considered ML algorithms implemented on datasets for stock market prediction. Moreover, ensemble learning approach argues about the best considerations and enhances the overall accuracy of model.