Classification of stock market index based on predictive fuzzy decision tree

Abstract

Over the past decade many attempts have been made to predict stock market data using statistical and data mining models. However, most methods suffer from serious drawback due to requiring long training times, results are often hard to understand, and producing inaccurate predictions. In addition, the trader’s expectations to predict stock markets are seriously affected by some uncertain factors including political situation, oil price, overall world situation, local stock markets etc. Therefore, predicting stock price movements is quite difficult. Data mining techniques are able to uncover hidden patterns and predict future trends and behaviors in financial markets. In this research, another modification of Fuzzy Decision Tree (FDT) classification techniques called predictive FDT is presented that aims to combine symbolic decision trees in data classification with approximate reasoning offered by fuzzy representation. The intent is to exploit complementary advantages of both: ability to learn from examples, high knowledge comprehensibility of decision trees, and the ability to deal with uncertain information of fuzzy representation. In particular, predictive FDT algorithm is based on the concept of degree of importance of attribute contributing to the classification. After constructing predictive FDT, Weighted Fuzzy Production Rules (WFPRs) are extracted from predictive FDT, and then more significant WFPR’s are mined by using similarity-based fuzzy reasoning method. In fuzzy reasoning method the weights are assigned to each proposition in the antecedent part and the Certainty Factor (CF) is computed for the consequent part of each Fuzzy Production Rule (FPR). Finally, these rules are used to predict time series stock market in different periods to time. The predictive FDT’s are tested using three data sets including Kuala Lumpur Stock Exchange (KLSE), New York Stock Exchange (NYSE) and London Stock Exchange (LSE). The experimental results show that the predictive FDT algorithm and fuzzy reasoning method provides the reasonable performance for comprehensibility (no of rules), complexity (no of nodes) and predictive accuracy of WFPRs for stock market time series data.