物流技术

城乡“客货邮公交”的服务质量水平直接关乎其可持续发展能力，因此有必要对其服务质量进行综合评估，进而提出具体的服务质量改进策略。文章从基础设施、运营管理、用户感知三个维度构建了城乡“客货邮公交”服务质量评价指标体系，运用机器学习方法分别建立了BP神经网络与随机森林两种基准评价模型，并通过引入主成分分析（Principal Component Analysis,PCA），构建了PCA-BP神经网络与PCA-随机森林两种综合评价模型。基于山东省莱西市城乡居民的调查问卷数据，对城乡“客货邮公交”服务质量进行了全面评价，并依据评价结果，提出了改进城乡“客货邮公交”服务质量的对策建议。研究发现：（1）PCA-随机森林模型表现出最优的预测性能；（2）无论是否引入PCA，随机森林模型的预测精度均优于BP神经网络模型；（3）经PCA改进后的机器学习评价模型在预测精度上均优于未经PCA处理的原始模型。

The service quality level of urban-rural "passenger-freight-postal bus" directly affects its sustainable development capability. Therefore, it is necessary to comprehensively evaluate its service quality and propose specific service quality improvement strategies. This paper constructs a service quality evaluation indicator system for urban-rural "passenger-freight-postal bus" from three dimensions: infrastructure, operational management, and user perception. Two baseline models are developed using machine learning methods, namely the backpropagation(BP) neural network and the random forest model. By introducing Principal Component Analysis(PCA), two comprehensive evaluation models, PCA-BP neural network and PCA-random forest, are further constructed. Based on questionnaire survey data from urban and rural residents in Laixi City, Shandong Province, a comprehensive evaluation of the service quality of the "passenger-freight-postal bus" is conducted.According to the evaluation results, countermeasures and suggestions for improving service quality are proposed. The results indicate that:(1) The PCA-random forest model demonstrates the best predictive performance.(2) Regardless of whether the PCA is introduced, the prediction accuracy of the random forest model is superior to that of the BP neural network model.(3) The machine learning evaluation models improved by PCA perform better than the original models without PCA processing in terms of prediction accuracy.