Stock Market Prediction Using ANN, SVM, ELM: A Review

Abstract Now-a-days stock market has become a major research area due to its non-linear behaviour in stock prices. So, a discerning prediction model is required to minimize high risk and maximize returns associated with sock prices. Several studies give solid demonstrations that models adapting traditional regression techniques face compelling challenges in out-of sample predictability test due to model ambiguity and parameter inconstancy. Soft computing techniques are tenable methods for considerable forecasting outcomes. This paper is a review of Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Extreme Learning Machines (ELM) attain and applied to predict stock prices. ANN is non- linear and nonparametric classifier which is viable for forecasting of stock prices. SVM uses the marginal values rather than average values for the classification prediction model. ELM uses fast training mechanism which is commercial for stock price prediction. Through this review it is unveil that these three data mining techniques are accustomed for studying and estimating stock market behavior.