CAJ | 학술논문

采用持向量机方法构建了居民出行方式选择模型,使用网格搜索方法选择支持向量参数,避免参数选择的随机性,分析不同核函数对模型构建的影响。研究表明,采用多项式核函数、RBF核函数构建的支持向量机模型对居民出行方式预测精度较高,所构建的模型可用于居民出行方式预测;在支持向量机核函数选择中,优先选择RBF核函数,其次为多项式核函数。
채용지향량궤방법구건료거민출행방식선택모형,사용망격수색방법선택지지향량삼수,피면삼수선택적수궤성,분석불동핵함수대모형구건적영향。연구표명,채용다항식핵함수、RBF핵함수구건적지지향량궤모형대거민출행방식예측정도교고,소구건적모형가용우거민출행방식예측;재지지향량궤핵함수선택중,우선선택RBF핵함수,기차위다항식핵함수。
This paper adopts support vector machine method to construct option model of residents＇ trip,and uses grid-search method to select parameters.This method avoids the randomness in choosing the parameters.Then it analyzes the influence of different kernel function to SVM model.The result shows that the support vector machines with polynomial kernel functions and RBF kernel functions obtain high accuracy in trip mode prediction.The constituted model can be used for residents＇ trip forecasting.In the choice of support vector machine kernel function,RBF kernel function is the privilege,polynomial kernel function is the followed.