A Comparative Analysis of Data Standardization Methods on Stock Movement

Abstract

Prediction of stock market indices has pinched considerable debate due to its brunt on economic development. Prediction of appropriate stock market indices is important in order to curtail the uncertainty related to it in order to arrive at conclusion on effective finance schemes. Thus selection of a proper forecasting model is highly appreciated which is always affected by the input data. The objective of this paper is to efficiently normalize input data in order to obtain accurate forecasting of stock movement and compare the accuracy results for different classifiers. This study compares three normalization techniques and their effect on the forecasting performance. In our work, we implemented different classifiers like SVM, ANN, and KNN for stock trend forecasting because of their risk management capabilities. This article deals primarily with the normalization of input data for the estimation of stock movement. Simulation was performed on six stock indices from different parts of the world market and a performance review of the system was performed. The study reveals the great affectability of commonly used methodologies for data standardization computations, as well as the need for a cautious approach to dealing with the results obtained.KeywordsArtificial neural networkInput data normalizationK nearest neighborPredictionSupport vector machinesStock market indices