Research On User Satisfaction Of New Energy Vehicles Based On Forum Comments

With the development and progress of digital technology,new energy vehicles have become an important part of the national development plan.However,my country’s new energy vehicle industry started late,and the product maturity is not high,which is easy to cause consumers’ anxiety and dissatisfaction,and has an impact on the development of the new energy vehicle industry.Consumer satisfaction surveys are an important means for companies to understand consumers’ preferences and demands,improve products,and enhance the competitiveness of car companies.The traditional satisfaction survey methods include questionnaires,telephone interviews,face-to-face interviews,etc.These survey methods have some limitations.For example,the questionnaire surveys are mostly designed from the perspective of enterprises,the research granularity is relatively coarse,and cannot reflect the detailed factors from the perspective of consumers.The research on the satisfaction of new energy vehicles in the literature is relatively scarce.Auto Forum is a place for car lovers to share their car buying experience.It provides real and rich data from the perspective of consumers and provides good data support for consumer satisfaction surveys.Therefore,in view of the shortcomings of existing research and the data characteristics of automobile forums,this study is based on the forum comment data,and uses text mining technology to quantify consumers’ satisfaction with new energy vehicles at different granularity levels.The main research contents are as follows :First,a three-level hierarchical consumer index system is constructed.Firstly,compound noun phrases are extracted from text reviews by combining syntactic dependencies and part-ofspeech tagging as an unregistered thesaurus in the field of new energy vehicles,and then the unregistered thesaurus is used to segment and tag the comments again,and filter out all the nouns and noun phrases as candidate evaluation indicators.At the same time,based on the largescale automobile forum corpus,the Word2 Vec tool is used to train word vectors.Finally,based on the candidate evaluation indicators and word vectors,the hierarchical clustering algorithm is used,supplemented by manual supervision,to construct a three-level hierarchical index system.Through the experimental verification of the data of BYD pure electric new energy vehicles,the result analysis proves the reliability of the research method.Second,an improved AATAE＿BiLSTM attribute-level sentiment analysis model is proposed,and based on the obtained hierarchical index system and sentiment tendency value,the satisfaction calculation of indicators at all levels is realized.Firstly,the AATAE＿Bi LSTM model is used to obtain the emotional tendencies of all the third-level indicators.Then,by writing an algorithm program,a comprehensive emotional score matrix of the third-level indicators is constructed.Finally,this study proposes a comprehensive calculation model of satisfaction.Based on the comprehensive emotional score matrix and the hierarchical index system,the calculation of the satisfaction and attention of the secondary and primary indicators is realized.Through the verification on the research data of this paper,the results show that the sentiment analysis model and the comprehensive satisfaction calculation model proposed in this study can quantify consumer satisfaction at both coarse-grained and fine-grained levels.For enterprises,this research can not only help them understand the current situation of products at the macro level,but also explore consumers’ consumption pain points and consumption preferences at the micro level,so as to improve and upgrade products and enhance product competitiveness.For consumers,their demands can be conveyed to enterprises,and it also broadens the channels for consumers to understand new energy vehicle consultation.