An empirical study of applying data mining techniques to the prediction of TAIEX Futures

It is an inevitable trend to learn and extract useful knowledge from massive data, so that data miming has been one of popular fields for researches and practitioners. Recently, data stream mining has emerged as an important subfield of data mining, because data samples usually are generated in a sequence over time and collected in a form of a stream in many cases in the real world. In this paper, we study a real-world problem and apply data stream mining techniques to the prediction of Taiwan Stock Exchange Capitalization Weighted Stock Index Futures (TAIEX Futures). We model the problem as a binary classification problem and our goal is to predict the rising or falling of the short-term futures. We design the data preprocessing procedure and employ a data stream miming toolkit in experiments. The results indicate that the concept drift detection method is helpful for TAIEX Futures in which concept drift supposedly exists and also that data stream mining technology is helpful for predicting the futures market.