Comparative Analysis of Machine Learning Models to Predict Stock Market Price

Abstract

AbstractStock market forecasting is viewed as one of the most interesting is of study for many researchers. The crucial data that can be accessed is thought to have predictive correlations with future stock performance which could provide information to investors so that they may make better decisions when purchasing equities. The paper tries to present a comparative analysis of four machine learning models to predict stock market price. The methods that we have considered are: support vector machine (SVM), artificial neural network (ANN), and hybrid models like PCA + SVM and PCA + ANN to predict stock market state. We have experimented using Vanguard Total Stock Market ETF (VTI) dataset for last 10 years which shows that SVM-based predictive model performed well among all the models for predicting the stock market status.KeywordsStock market predictionPrincipal component analysisMachine learningANNSVMHybrid models