Explorative data mining on stock data -experimental results and findings

Abstract

This paper presents our preliminary explorative data mining experiments on stock data. The experiments are performed using information gain evaluation, chi-square test and decision tree, which aim to explore the underlying patterns of the stock dataset. The explorative patterns obtained show some unexpected outliers in deriving the influential stocks with regard to Kuala Lumpur Composite Index. © Springer-Verlag Berlin Heidelberg 2007.