A hybrid approach by integrating wavelet-based feature extraction with MARS and SVR for stock index forecasting

Authors:
Ling-Jing Kao;Chih-Chou Chiu;Chi-Jie Lu;Chih-Hsiang Chang
Affiliations:
Department of Business Management, National Taipei University of Technology, Taiwan;Department of Business Management, National Taipei University of Technology, Taiwan;Department of Industrial Management, Chien Hsin University of Science and Technology, Taiwan;Institute of Commerce Automation and Management, National Taipei University of Technology, Taiwan
Venue:
Decision Support Systems
Year:
2013

Citing 22
Cited 3

A Theory for Multiresolution Signal Decomposition: The Wavelet Representation

IEEE Transactions on Pattern Analysis and Machine Intelligence
The nature of statistical learning theory

The nature of statistical learning theory
A Stochastic Nonlinear Regression Estimator Using Wavelets

Computational Economics
A survey on wavelet applications in data mining

ACM SIGKDD Explorations Newsletter
Practical selection of SVM parameters and noise estimation for SVM regression

Neural Networks
Predicting object-oriented software maintainability using multivariate adaptive regression splines

Journal of Systems and Software
Financial Time Series Forecasting Using a Compound Model Based on Wavelet Frame and Support Vector Regression

ICNC '08 Proceedings of the 2008 Fourth International Conference on Natural Computation - Volume 05
Prediction Model of Stock Market Returns Based on Wavelet Neural Network

PACIIA '08 Proceedings of the 2008 IEEE Pacific-Asia Workshop on Computational Intelligence and Industrial Application - Volume 01
Surveying stock market forecasting techniques - Part II: Soft computing methods

Expert Systems with Applications: An International Journal
Financial time series forecasting using independent component analysis and support vector regression

Decision Support Systems
Forecasting stock market short-term trends using a neuro-fuzzy based methodology

Expert Systems with Applications: An International Journal
An Adaptive Network-Based Fuzzy Inference System (ANFIS) for the prediction of stock market return: The case of the Istanbul Stock Exchange

Expert Systems with Applications: An International Journal
A hybrid ANFIS model based on AR and volatility for TAIEX forecasting

Applied Soft Computing
Combining multiple feature selection methods for stock prediction: Union, intersection, and multi-intersection approaches

Decision Support Systems
A hybrid SARIMA wavelet transform method for sales forecasting

Decision Support Systems
Multiresolution FIR neural-network-based learning algorithm applied to network traffic prediction

IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
A Hybrid System Integrating a Wavelet and TSK Fuzzy Rules for Stock Price Forecasting

IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
Predicting stock market returns from malicious attacks: A comparative analysis of vector autoregression and time-delayed neural networks

Decision Support Systems
Texture classification and segmentation using wavelet frames

IEEE Transactions on Image Processing
Multiresolution forecasting for futures trading using wavelet decompositions

IEEE Transactions on Neural Networks
Support vector machine with adaptive parameters in financial time series forecasting

IEEE Transactions on Neural Networks
Sales forecasting for computer wholesalers: A comparison of multivariate adaptive regression splines and artificial neural networks

Decision Support Systems

Sales forecasting of computer products based on variable selection scheme and support vector regression

Neurocomputing
A hybrid WA-CPSO-LSSVR model for dissolved oxygen content prediction in crab culture

Engineering Applications of Artificial Intelligence
A real time clustering and SVM based price-volatility prediction for optimal trading strategy

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Forecasting stock prices is a major activity of financial firms and private investors when they make investment decisions. Feature extraction is usually the first step of a stock price forecasting model development. Wavelet transform, used mainly for the extraction of information contained in signals, is a signal processing technique that can simultaneously analyze the time domain and the frequency domain. When wavelet transform is employed to construct a forecasting model, the wavelet basis functions and decomposition stages need to be determined first. However, because forecasting models constructed by different wavelet sub-series would exhibit different forecasting capabilities and yield varying forecast results, the selection of wavelet that can lead to an optimal forecast outcome is extremely critical in model construction. In this study, a new stock price forecasting model which integrates wavelet transform, multivariate adaptive regression splines (MARS), and support vector regression (SVR) (called Wavelet-MARS-SVR) is proposed to not only address the problem of wavelet sub-series selection but also improve the forecast accuracy. The performance of the proposed method is evaluated by comparing the forecasting results of Wavelet-MARS-SVR with the ones made by other five competing approaches (Wavelet-SVR, Wavelet-MARS, single ARIMA, single SVR and single ANFIS) on the stock price data of two newly emerging stock markets and two mature stock markets. The empirical study shows that the proposed approach can not only solve the problem of wavelet sub-series selection but also outperform other competing models. Moreover, according to the sub-series which are selected by the proposed approach, we can successfully identify the data of which sessions (or points in time) among past stock market prices exerted significant impact on the construction of the forecasting model.