摘要
为更好地把握春运期间广大旅客出行规律,深入了解旅客出行体验和服务需求,春运期间旅客在出行过程中对某些热点问题的关注度,以及对某种交通出行方式的满意度及情感倾向的分析与研究是一个很好地切入点。首先以实际采集的旅客在春运调查问卷中的开放性意见文本数据为基础,采用基于语义相似度和词向量方法,并结合机器学习算法分别构建了基础情感词典和交通运输领域情感词典。其次,在此基础上结合语义规则,基于春运交通大数据构建了运输领域旅客情感分析模型。最后,通过对实际采集的春运期间针对旅客出行的开放性问卷调查回收海量数据进行测试,验证了所提出的分析策略和模型能有效分析识别旅客的情感信息。其中,基础数据整合了春运期间铁路、航空、水运、公路等多种出行方式的旅客开放性意见海量数据,数据具有广泛的覆盖性和较强的代表性。从分析结果可以看出,北京、广东、江苏和四川等主要省市的旅客对购票问题更为关注,占比在30%左右;相比而言,河南和山东的旅客更为关注安全和服务,占比达到了33%以上。结果与实际情况是相符的,进一步证明所提出的模型具有很强的适应性,其结果可以作为交通运营服务评价和交通管理决策的重要依据。
In order to better grasp the travel rules of the vast number of passengers during e Spring Festival travel period, and deeply understand the travel experience and service demand of passengers, the analysis and research on the passengers’ attention to some hot issues as well as the satisfaction and emotional tendency of a certain mode of transport during the Spring Festival travel period are the good breakthrough point. First, based on the actual collected open opinion text data of passengers in the Spring Festival transport questionnaire, by using semantic similarity and word vector based method, and combining with the machine learning algorithm, the basic emotion dictionary and the emotion dictionary in the transport field are constructed respectively. Second, combining with the semantic rules, the passenger emotion analysis model in the transport field is built based on the big data of Spring Festival travel rush accordingly. Finally, by carrying out the test of the actual collected big data of open questionnaire survey of Spring Festival travel rush of passengers, it is verified that the analysis strategy and proposed model can effectively analyze and identify the emotional information of passengers. Among them, the basic data integrates a large amount of open opinion data of passengers from railway, aviation, water transport, highway and other travel modes during the Spring Festival transport period. The data has a wide coverage and strong representativeness. The analysis result shows that(1) passengers in Beijing, Guangdong, Jiangsu and Sichuan are more concerned about ticket purchase, accounting for about 30%;(2)in contrast, travelers from Henan and Shandong pay more attention to safety and service, accounting for more than 33%. The result is in accordance with the actual situation, so the proposed model has a strong adaptability, the result can be used as an important basis for traffic operation service evaluation and traffic management decision making.
作者
郭宇奇
查文斌
李斌
刘冬梅
GUO Yu-qi;ZHA Wen-bin;LI Bin;LIU Dong-mei(Research Institute of Highway,Ministry of Transport,Beijing 100088,China)
出处
《公路交通科技》
CAS
CSCD
北大核心
2020年第3期105-113,共9页
Journal of Highway and Transportation Research and Development
基金
中央公益性科研院所基本科研业务费专项资金项目(2019-0092,2019-0019,zx-2017-014)
“十三五”国家重点研发计划项目(2017YFC0840206)。
关键词
智能交通
旅客情感分析
机器学习
春运大数据
词向量
ITS
passenger emotion analysis
machine learning
big data of spring festival travel rush
word vector
