Research On Evaluation Of New Energy Vehicle’s Reputation Based On Text Mining

In order to alleviate the problems of rapid urbanization and energy consumption,the development of new energy vehicles has received strong support from the state,and it will certainly become the direction and goal of the future development of the automotive industry.Although the new energy automobile industry is developing rapidly today,it is still in the groping stage in the mainstream automobile market.In order to better increase the enthusiasm for consumption in the new energy vehicle market,it is of great practical significance to explore and analyze people’s views or use experience of new energy vehicles.This article discusses the new energy vehicle word-of-mouth comment data on Autohome and the website of Daochedi as the research object.First,use the web crawler technology to obtain the required word-of-mouth data,classify according to the price of the car and the difference of the model,and classify the current situation of the new energy vehicle.Perform descriptive statistical analysis.Then carry out data cleaning and word segmentation processing on the text.In order to be able to segment words more accurately,this thesis constructs a dictionary of automobile proprietary names.Use word cloud graphs and semantic networks to analyze the characteristics of the review text and visualize the user’s focus on new energy vehicles.Then,the satisfaction dimension in the word-of-mouth comment data is marked as positive sentiment text,and the dissatisfied dimension is marked as negative sentiment text.The machine learning model Naive Bayes,XGBoost,SGD,and the deep learning model LSTM are used for model training on the comment text that has been marked,and the model is verified with the test set.By comparing the effects of the four sentiment classifiers,the best LSTM classifier was selected to label unlabeled comments.Then the LDA theme feature extraction is performed on the interior and cost-effective dimensions of the new energy vehicle,and on this basis,the advantages and disadvantages of the car are analyzed by the CorEx theme model with anchored vocabulary to dig out consumers’ focus on new energy vehicles.Finally,summarize the research results of this article and point out further issues that need to be considered.The main conclusions obtained from the research are as follows: First,the main market for new energy vehicles is economically developed cities,and SUVs and sedans of 100,000 to 200,000 are the main models in the new energy vehicle consumer market.Second,in the sentiment classification task of word-of-mouth reviews of new energy vehicles,the deep learning model LSTM has a prediction accuracy of 97.43% in the sentiment classification of review text,which is higher than the machine learning models Naive Bayes,XGBoost and SGD.Third,the advantages of new energy vehicles at this stage are sufficient power,low energy consumption and good appearance.Consumers’ main dissatisfaction with new energy vehicles is manifested in software such as interiors and comfort.Combining the above analysis,we put forward targeted suggestions: The government needs to strengthen the construction of charging infrastructure for new energy vehicles to solve the problem of regional development imbalance.Manufacturers should pay attention to consumer demand.In addition to paying attention to the battery life of new energy vehicles,they should also pay attention to the consumer’s ride experience.