Predicting Direction of Stock Index Movement by Unification of Two Stage Machine Learning Models

Abstract

The paper proposes a two stage unification of machine learning models to predict the direction of next day stock index movement. The study is conducted using four prediction models – artificial neural network, support vector machine, random forest and naive bayes. Stage one uses the direction classification that is either up or down by incorporating signals from five technical indicators. Indicators are calculated as per their definition by using daily trading data of open, high, low, close and volume data. Stage two takes in the processed discretised data for predictions from stage one. The accuracy of both stages for each model is evaluated and compared. Experimental results show significant improvement of the combined model over single-pass model. Artificial neural network provided the best accuracies which is closely followed by support vector and random forest.