支持向量机在出行链模式识别和影响因素分析中的应用.pdf

Journal of Southeast University (English Edition) Vol. 33, No. 1, pp. 106 — 114 Mar. 2017 ISSN 1003—7985 Application of support vector machine in trip chaining pattern recognition and analysis of explanatory variable effects Yang Shuo Deng Wei Cheng Long (School of Transportation, Southeast University, Nanjing 210096, China) Abstract: In order to improve the accuracy of travel demand trip chaining behavior, and the model results show that forecast and considering the distribution of travel behaviors activity decision or trip chaining behavior is made before within time dimension, a trip chaining pattern recognition mode choice in most cases. Meanwhile, household char­ model was established based on activity purposes by applying acteristics and personal attributes ae also found signifi­ three methods: the support vector machine (SVM) model, the cantly infuencing trip chaining behavior, such as house­ radial basis function neural network (RBFNN) model and the hold members5 work status[3], work schedule[4], gen- multinomial logit (MNL) model. The effect of explanatory der[5], the presence of children[6], etc. In addition, the factors on trip chaining behaviors and their contribution to built environment where people live and work is another model performace were investigated by sensitivity analysis. important factor, and is effects on trip chaining behaviors Results show that the SVM model has a better performance have found their way into the diverse concepts of plan­ than the RBFNN model and the MNL model due to its higher ning, design and policies[779, such as a compact city, overa