International Conference of China Communication and Technology (ICCCT 2010 E-BOOK)

Nan'ning,China,10.9-10.10,2010

ISBN: 978-1-935068-52-5 Scientific Research Publishing, USA

E-Book 454pp Pub. Date: October 2010

Category: Computer Science & Communications

Price: $80

Title: The TAR Forecast Model-Based on the Data Set D Establishment
Source: International Conference of China Communication and Technology (ICCCT 2010 E-BOOK) (pp 173-176)
Author(s): Xiang-kun Xu, Information Department, Polytechnic School of Shenyang Ligong University, Fushun 113122, China
Hua-bing Xu, Department of Dynamic Engineering, Bohai Shipbuilding Vocational College, Huludao 125000, China
Abstract: To further improve the accuracy of nonlinear time series prediction, a robust threshold auto regression model used to study data set D is developed. The delay characteristics of the threshold autoregressive prediction model for the time series is studied by Self-correlation analysis techniques. The point value chart (Dot Figure) is used to determine intervals number of the TAR threshold and threshold optimization scope. Furthermore, randomized algorithms have also been employed to optimize the threshold value and the regression coefficient of the TAR model. The experimental results indicate that the established threshold regression model with relative low error can accurately predict the nonlinear time series.
Free SCIRP Newsletters
Copyright © 2006-2024 Scientific Research Publishing Inc. All Rights Reserved.
Top