基于社交网络数据的城市轨道交通服务质量评价模型

Evaluation model of urban rail transit service quality based on social network data

服务质量中乘客感知的获取和满足是提高城市轨道交通吸引力的重要依据,评价过程的科学性与结果的准确性将对优化城市轨道交通的运营管理产生关键影响。为解决城市轨道交通服务质量主要依靠问卷调查,无法全面反映乘客真实心理感知的问题,以社交网络评论数据为切入点,运用自然语言处理技术,对轨道交通服务质量评价进行量化研究。首先,通过网络爬虫技术对社交网络中相关评论数据进行采集,针对文本预处理结果,运用基于情感词典构建与量级划分的分析方法,识别语料情感极性和强度。然后,建立基于K-Means文本聚类算法的轨道交通服务质量评价指标体系,将乘客需求与服务要素转化为评价指标,应用TF-IDF法,结合文本特征评估指标重要度,计算服务质量综合评价得分。最后,选取微博平台中重庆轨道交通评论语料为例进行实证分析。研究结果表明:重庆轨道交通服务质量综合评价分值为4.383,总体处于较低水平,运营服务提升空间较大;乘客对检票智能及人员服务方面满意度最高,车厢温度情感得分最低;影响服务质量最重要的因素为乘车安全(7.850%),其次分别是票价经济(7.524%)、购票便捷(7.212%)和检票智能(7.139%)。相较现有方法,社交网络数据可更为直观地反馈乘客意见,为轨道交通服务质量评价提供科学的数据来源。

The acquisition and satisfaction of passengers’perception is the key basis for improving the attractiveness of urban rail transit.The scientific evaluation process and accurate results will bring critical impacts to the optimization of urban rail transit operation management.To solve the problem that the absolute dependence of urban rail transit service quality on questionnaire surveys cannot reflect the real psychological perception of passengers,this paper used natural language processing to quantify the service quality of rail transit based on social network data.First,a web crawler was used to collect comment data in social networks,and an analysis method based on sentiment lexicon construction and degree division was adopted for determining the sentiment polarity and intensity of the text preprocessing results.Then,by establishing the evaluation index system of rail transit service quality based on K-Means text clustering,passengers’demands and service elements were transformed into evaluation indexes.The index weight was acquired by TF-IDF calculation of the importance of feature items,and the comprehensive score of service quality was obtained by combining the results.Finally,the corpus of Chongqing rail transit comments in Microblog was selected as the case for empirical analysis.The results prove that the service quality of Chongqing rail transit scores 4.383,which is generally at a low level.There is still much room for improvement in operation services.Passengers have the highest satisfaction with the intelligence of ticket checking and personnel services,and the lowest emotional score on compartment temperature.The key factor affecting service quality is safety(7.850%),followed by economical fares(7.524%),ticket convenience(7.212%),and intelligent ticket checking(7.139%).Compared with the existing methods,social network data can more intuitively identify passengers’opinions,providing a scientific data source for the evaluation of rail transit service quality.