Title :
Gross industrial output value prediction based on least squares support vector regression
Author_Institution :
Econ. & Manage. Sch., Wuhan Univ., Wuhan, China
Abstract :
Least squares support vector regression is presented in gross industrial output value prediction in the paper. Least squares support vector regression is a kind modified support vector regression. It can solve a convex quadratic programming problem, which has higher performance than support vector regression. The data of gross industrial output value in Fujian province from 1990 to 2006 are employed to train and test the proposed model. It is indicated that prediction performance of gross industrial output value of LSSVR model is best in the RBFNN, SVR and LSSVR prediction model. Then, LSSVR has very high application values in prediction of gross industrial output value.
Keywords :
convex programming; forecasting theory; industrial economics; least squares approximations; logistics; quadratic programming; radial basis function networks; regression analysis; support vector machines; Fujian province; RBFNN; convex quadratic programming; gross industrial output value prediction; least square method; support vector regression; Artificial neural networks; Data models; Forecasting; Predictive models; Support vector machines; Testing; Training; Least squares support vector regression; gross industrial output value; prediction performance;
Conference_Titel :
Signal Processing Systems (ICSPS), 2010 2nd International Conference on
Conference_Location :
Dalian
Print_ISBN :
978-1-4244-6892-8
Electronic_ISBN :
978-1-4244-6893-5
DOI :
10.1109/ICSPS.2010.5555821
