论文结合我国公共交通发展特点确定48个候选指标.运用德尔菲法进行指标初选:在第一轮调查中,以关联程度赋值平均值大于3.0、变异系数小于0.25为筛选标准删除8个指标;在第二轮计算后,采用BP神经网络法对指标进行优化,以权重小于60%为筛选标准删除11个指标,最终筛选出29个指标.又将评价体系分为目标层、准则层和指标层;通过训练神经网络,得到各指标权重及准则层权重.最后,以南京为例进行指标体系的应用,应用结果表明,该指标体系能够较准确和全面评价公共交通发展水平.%The paper determines 48 candidate indicators based on the characteristics of public transport in our country.In the first round of surveys,the average value of assignment was greater than 3.0 and the coefficient of variation was less than 0.25.Eight indicators were deleted by using the Delphi method.After the second round of calculation,BP neural network indicators were optimized,with the weight less than 60% as the screening criteria to delete 11 indicators,and the final selection was 29 indicators.The evaluation system is divided into the target layer,the guideline layer and the indicator layer.By training the neural network,the weight of each indicator and the weight of the guideline layer are obtained.Finally,taking Nanjing as an example,the application of index system is made.The application results show that this index system can evaluate the level of public transport more accurately and comprehensively.
展开▼
