基于BP神经网络的城市公交服务质量影响因素主成分分析

Principal Component Analysis of Factors Influencing City Bus Service Quality Based on BP Neural Network

为改善城市公交服务质量,根据乘客服务质量问卷调查数据分析,运用主成分分析方法对公交服务质量影响因素进行降维处理,把15个影响变量提取为8个主成分变量。在此基础上,运用MATLAB7.0建立影响因素主成分与公交服务质量认可度的BP神经网络模型,在不同参数下进行试验和比较,计算出模拟数据下公交服务质量认可度折减影响系数均方差为0.000 957,表明该模型所选参数值可以用于评估公交服务质量的影响因素分析。最后,根据权值和阀值计算得出影响城市公交服务质量的关键因素为公交车内拥挤程度、驾驶员服务态度和首末班车时间,其影响程度分别为53.09%、32.02%和30.36%。研究结论可以为城市公交服务质量改善提供依据并明确重点改进的方向。

In order to improve city bus service quality, the factors dimension of bus service quality werereduced by principal component analysis and 8 principal components were extracted from 15 variablesaccording to the passenger service quality survey data analysis. Then, the BP neural network modelwhich reflected the relations about the factors principal component and recognition was set up by usingMATLAB7.0. By testing and comparing with different parameters, the weight matrix was calculated.Based on simulated data, it was calculated that the bus service quality recognition reduction impact coef.ficient variance was 0.000957, which indicated that the parameter values selected in model were usedfor the factors analysis to assess the bus service quality. Finally, according to weights and thresholds, itis calculated that the main factors influencing the bus service quality were passengers crowding, driverattitude and the first and last time of bus, whose impact were 53.09%, 32.02% and 30.36%. The conclu.sion provided the basis and the clear direction for the city bus service quality improvement.