Int. J. of Data Science   »   2016 Vol.1, No.3



Title: Empirical mode decomposition-based models for predicting direction of stock index movement


Authors: Youqin Pan; Yong Hu


Department of Marketing and Decision Science, Salem State University, 352 Lafayette Street, Salem, MA 01970, USA
Institute of Business Intelligence and Knowledge Discovery, Guangdong University of Foreign Studies, Guangzhou Higher Education Mega Center, Guangzhou 510006, China


Abstract: In this paper, novel forecasting models based on empirical mode decomposition (EMD) are proposed to predict the direction of stock market movement. The proposed models first use EMD to adaptively decompose the complicated stock index into a small number of intrinsic mode functions (IMFs). Then, these IMFs were used as explanatory variables to predict the signs of stock market movement. The Dow Jones industrial average (DJIA) index, Hang Seng index (HSI) and Shanghai stock exchange composite (SSE) index were used to evaluate the performance of the proposed models. The proposed learning algorithms generate about a 70% hit ratio on weekly stock indices except that of the logistic regression on SSE. Moreover, the learning algorithms seem to perform equally well on the monthly stock indices and the weekly Dow index. However, there are significant differences among the model performances of the three learning algorithms on weekly SSE and HSI indices.


Keywords: EMD; empirical mode decomposition; logistic regression; SVM; support vector machines; ANNs; artificial neural networks; prediction modelling; stock index movement; stock markets; forecasting models; model performance.


DOI: 10.1504/IJDS.2016.075936


Int. J. of Data Science, 2016 Vol.1, No.3, pp.205 - 226


Available online: 12 Apr 2016



Editors Full text accessPurchase this articleComment on this article