Genetic algorithms approach to feature discretization in artificial neural networks for the prediction of stock price index

Abstract

This paper proposes genetic algorithms (GAs) approach to feature discretization and the determination of connection weights for artificial neural networks (ANNs) to predict the stock price index. Previous research proposed many hybrid models of ANN and GA for the method of training the network, feature subset selection, and topology optimization. In most of these studies, however, GA is only used to improve the learning algorithm itself. In this study, GA is employed not only to improve the learning algorithm, but also to reduce the complexity in feature space. GA optimizes simultaneously the connection weights between layers and the thresholds for feature discretization. The genetically evolved weights mitigate the well-known limitations of the gradient descent algorithm. In addition, globally searched feature discretization reduces the dimensionality of the feature space and eliminates irrelevant factors. Experimental results show that GA approach to the feature discretization model outperforms the other two conventional models.

Introduction

For a long time, there has been much research interest in predicting the stock price index. Among them, there are many studies using data mining techniques including artificial neural networks (ANNs). However, most studies showed that ANN had some limitations in learning the patterns because stock market data has tremendous noise and complex dimensionality. ANN has preeminent learning ability while it is often confronted with inconsistent and unpredictable performance for noisy data. In addition, sometimes the amount of data is so large that the learning of patterns may not work well. In particular, the existence of continuous data and large amount of data may pose a challenging task to explicit concepts extraction from the raw data due to the huge amount of data space determined by continuous features (Liu & Setiono, 1996). Many researchers in the society of data mining are interested in the reduction of dimensionality. The reduction and transformation of the irrelevant or redundant features may shorten the running time and yield more generalized results (Dash & Liu, 1997).

This paper proposes a new hybrid model of ANN and genetic algorithms (GAs) for feature discretization to mitigate the above limitations. Feature discretization is to transform continuous values into discrete ones in accordance with certain thresholds. Feature discretization is closely related to the dimensionality reduction (Liu & Motoda, 1998a). Properly discretized data can simplify the process of learning and may improve the generalizability of the learned results. This study uses GA to search the optimal or near-optimal thresholds for feature discretization. In addition, this study simultaneously searches the connection weights between layers in ANN. The genetically evolved connection weights mitigate the well-known limitations of the gradient descent algorithm.

The rest of the paper is organized as follows. Section 2 reviews prior research. Section 3 proposes feature discretization using GA and describes the benefits of the proposed approach. Section 4 describes the research design and experiments. In Section 5, the empirical results are summarized and discussed. In Section 6, conclusions and the limitations of this study are presented.