Online Comment Opinion Mining And Econometric Analysis Based On Deep Learning

There are many kinds of word-of-mouth information in online comments,and their emotional tendency has a great influence on consumers' repurchase intention.Opinion mining of online comments refers to the analysis of the emotional orientation of different object types from the comment text.Generally,the emotional orientation of the evaluation object emphasized by consumers has a greater impact on the overall emotional orientation of the comment.The types of evaluation objects can be divided into explicit evaluation objects and implicit evaluation objects.In the current research,more attention is paid to explicit evaluation objects in the text,while the implicit evaluation objects in the text are ignored.The online word-of-mouth information included in the reviews can be divided into commodity quality and merchant service.Currently,few studies have studied the influence of commodity quality and merchant service emotional polarity on consumers' purchase intention,as well as the influence relationship between commodity quality,merchant service and consumers' emotional score of repurchase intention.In this paper,the deep learning model is optimized in the analysis of online comment affective orientation.Firstly,the key evaluation objects in the text are extracted according to the semantic information of the text.In order to improve the performance of model classification,a BiLSTM model is constructed which integrates key object recognition and deep attention mechanism.Based on the splicing of key evaluation objects in the input layer of the BiLSTM model,the deep attention mechanism is used in both the input and output layers of the model.The model constructed was tested with the hotel review data set,and the experimental results showed that the model constructed in this paper had the best classification effect.The precision rate,recall rate and F1 value of the model were 93.01%,93.02% and 93.01% respectively.After using the constructed model of online sales review data of xinjiang characteristic agricultural products are analyzed,and each of the online comments tend to have more than one evaluation dimensions,this article to comment on each of the text respectively from three dimensions(commodity quality,business services,consumer purchase intention)emotional score was calculated,and according to the score of each dimension text respectively from high praise,medium review and bad review three dimension emotion classification.In the study of consumers' repurchase intention,PVAR(1)model was firstly used to analyze the relationship between commodity quality,merchant service and consumers' emotional score of consumers' repurchase intention.The empirical results show that the affective score of goods quality and merchant service lags behind the first order and has a positive effect on the emotional score of consumers' repurchase intention.In order to improve consumers' willingness to buy again,merchants must provide satisfactory product quality and services.With panel then the Logistic regression model analyzes the factors that affect consumers' willingness to buy again,the empirical results show that both in taobao shop and flagship store positive evaluation of goods quality and business services and consumer purchase intention lag item again again will be positive influence on consumer purchase intention,and commodity quality and business service cross terms will boost consumer purchase intention.The regression results of the control variables(merchants' reputation level,sales volume,etc.)are consistent with the hypothesis,while the increase of promotion activities and comments will have a positive impact on the users of flagship stores but have no impact on taobao users.According to the research conclusion of this paper,the following Suggestions are proposed :(1)when classifying the text,more attention should be paid to the key evaluation objects in the text.(2)businesses should encourage consumers to use more explicit evaluation objects and provide comments from multiple dimensions.(3)in order to retain existing consumers,merchants should provide more high-quality goods and good services while maintaining the reputation level and credit score of their stores.It is not an inevitable measure to promote consumers' willingness to buy again in the sales process.