基于IPSO-XGB的城市轨道交通车站服务质量评价

Quality evaluation of urban rail transit station based on IPSO-XGB

针对城市轨道交通车站服务质量评价体系不够完善和指标赋权方法较单一的现状,结合当前乘客消费理念及出行习惯,构建基于乘客感知的初始评价指标体系。应用改进粒子群算法(IPSO)与极端梯度提升树(XGB)的混合算法IPSO-XGB计算各指标权重,结合乘客满意度,形成IPA矩阵,得到最终的评价结果,并进一步精简评价指标体系。以长沙地铁五一广场站为例,用分类误差率衡量算法优劣,对车站服务质量进行评价,并将其结果与分类回归树、神经网络等评价算法进行对比。研究结果表明:本文提出的IPSO-XGB评价算法的分类误差率最小,可降至3.85%。

In view of the imperfect evaluation system of urban rail transit station service quality and the single way of index weighting,combined with current passenger consumption concepts and travel habits,this paper proposed an initial evaluation index system based on passenger perception, and then calculated the weight of each indicator by using the hybrid algorithm of improved particle swarm optimization (IPSO) and extreme gradient boosting tree (XGB), named IPSO-XGB. In combination with passenger satisfaction, an IPA matrix was formed to obtain the final evaluation results and further simplify the evaluation index system. Taking the Wuyi Square Station of Changsha Metro as an example, the classification error rate was used to measure the pros and cons of the algorithm, evaluated the service quality of this station, and compared the results with the classification regression tree and neural network. The experimental results show that the classification error rate of the IPSO-XGB evaluation algorithm proposed in this paper is the smallest, which can be reduced to 3.85%.