Bias effect on predicting market trends with EMD

Ensemble Empirical Mode Decomposition is used for trend prediction of market indexes.Using EMD as a pre-processing step is shown to add look-ahead bias.We designed a protocol that eliminates look-ahead bias.8 market indexes and 4 different models were tested as part of the study.In contrast with published results, EEMD did not prove to be advantageous in general. Financial time series are notoriously difficult to analyze and predict, given their non-stationary, highly oscillatory nature. In this study, we evaluate the effectiveness of the Ensemble Empirical Mode Decomposition (EEMD), the ensemble version of Empirical Mode Decomposition (EMD), at generating a representation for market indexes that improves trend prediction. Our results suggest that the promising results reported using EEMD on financial time series were obtained by inadvertently adding look-ahead bias to the testing protocol via pre-processing the entire series with EMD, which affects predictive results. In contrast to conclusions found in the literature, our results indicate that the application of EMD and EEMD with the objective of generating a better representation for financial time series is not sufficient to improve the accuracy or cumulative return obtained by the models used in this study.

[1]  Yaguo Lei,et al.  A review on empirical mode decomposition in fault diagnosis of rotating machinery , 2013 .

[2]  Ganesh Mani,et al.  Financial Forecasting Using Genetic Algorithms , 1996, Appl. Artif. Intell..

[3]  Zhifang He,et al.  Stock Price Prediction based on SSA and SVM , 2014, ITQM.

[4]  Kyoung-jae Kim,et al.  Financial time series forecasting using support vector machines , 2003, Neurocomputing.

[5]  Saeid Sanei,et al.  Extracting fetal heart signal from noisy maternal ECG by singular spectrum analysis , 2010 .

[6]  T. Wang,et al.  Comparing the applications of EMD and EEMD on time-frequency analysis of seismic signal , 2012 .

[7]  Saeid Sanei,et al.  Extracting fetal heart signal from noisy maternal ECG by multivariate singular spectrum analysis , 2010 .

[8]  N. Huang,et al.  The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis , 1998, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences.

[9]  Rui Zhang,et al.  Hybrid Decomposition and Ensemble Framework for Stock Price Forecasting: a Comparative Study , 2011, Adv. Data Sci. Adapt. Anal..

[10]  Norden E. Huang,et al.  Ensemble Empirical Mode Decomposition: a Noise-Assisted Data Analysis Method , 2009, Adv. Data Sci. Adapt. Anal..

[11]  Gong Xu,et al.  Stock Price Prediction based on SSA and SVM , 2014 .

[12]  Stefan Lessmann,et al.  Bridging the divide in financial market forecasting: machine learners vs. financial economists , 2016, Expert Syst. Appl..

[13]  R. Vautard,et al.  Singular-spectrum analysis: a toolkit for short, noisy chaotic signals , 1992 .

[14]  Ravi Sankar,et al.  Time Series Prediction Using Support Vector Machines: A Survey , 2009, IEEE Computational Intelligence Magazine.

[15]  Oguz Akbilgic,et al.  A novel Hybrid RBF Neural Networks model as a forecaster , 2013, Statistics and Computing.

[16]  K. Lai,et al.  Forecasting crude oil price with an EMD-based neural network ensemble learning paradigm , 2008 .

[17]  S. S. Shen,et al.  A confidence limit for the empirical mode decomposition and Hilbert spectral analysis , 2003, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences.

[18]  Manish Kumar,et al.  Forecasting Stock Index Movement: A Comparison of Support Vector Machines and Random Forest , 2006 .

[19]  P. J. J. Luukko,et al.  Introducing libeemd: a program package for performing the ensemble empirical mode decomposition , 2015, Computational Statistics.

[20]  Maysam Abbod,et al.  A novel hybrid ensemble model to predict FTSE100 index by combining neural network and EEMD , 2015, 2015 European Control Conference (ECC).

[21]  T. Mikosch,et al.  Nonstationarities in Financial Time Series, the Long-Range Dependence, and the IGARCH Effects , 2004, Review of Economics and Statistics.

[22]  Ling Tang,et al.  EEMD-LSSVR-Based Decomposition-and-Ensemble Methodology with Application to Nuclear Energy Consumption Forecasting , 2011, 2011 Fourth International Joint Conference on Computational Sciences and Optimization.

[23]  Shichang Du,et al.  An Optimal Ensemble Empirical Mode Decomposition Method for Vibration Signal Decomposition , 2017 .