Prediction of Highway's Passenger Traffic Based on KPCA-SVM

HU Yan-rong,WU Chong,LIU Hong-jiu
DOI: https://doi.org/10.3969/j.issn.1004-292x.2012.01.002
2012-01-01
Abstract:Support vector machine is a machine learning method based on statistical learning theory and structural risk minimization.Support vector machine is much better method than ever,because the method may solve some actual problem in the small samples,high dimension,nonlinear and local minima and so on.The article utilizes the theory and method of support vector machine(SVM)regression,and establishes the regressive model based on Kernel Principal Component Analysis-Support Vector Machine(KPCA-SVM).Through forecasting the passenger traffic of Hangzhou's highway in 2000-2008,we show that the regressive model of KPCA-SVM has much higher accuracy and reliability of prediction,and may effectively forecast the passenger traffic of highway.
What problem does this paper attempt to address?