Analysis On Hot Topics Of Ride-hailing Based On Weibo Comments

With the rise of the "Internet +" model,the network car has been generated with the market demand,and the scale of the network car market has gradually expanded,which greatly facilitates people's travel and plays a role in rationally allocating market resources.However,as the scale expands,as more and more platforms join the market and compete with each other,major platforms have taken measures to attract users,compete for market resources,and seize market share,thus causing a series of problems.The pursuit of contention for the user community is only for the sake of its own platform to stand firm in this market,ignoring the user travel experience,resulting in many platforms can not continue to develop,and even face bankruptcy.In addition,the network car platform itself is not perfect,and many loopholes have caused many social problems.At present,there are few studies on the network car,and the form is single,mostly in the form of questionnaires.This paper takes advantage of the timeliness and topicality of Weibo,selects microblog commentary data for analysis,and combines sentiment analysis and topic mining to study the concerns of consumers in the development of online car.This paper collects the comment data generated by the US group taxis entering the market on the Weibo platform,firstly samples the network microblog commentary dataset,and extracts the new dataset after emoji extraction,subjective and objective text recognition and rule filtering.A total of 30,497 pieces of data were selected by support vector machine and convolutional neural network to classify the data of the review data and carry out model training.The parameters were optimized according to the training effect and the effect of the model was compared.Then based on the results of sentiment classification,the topic hotspot analysis,mining the overall comments and the hotspots of different emotional category comments,and analyzing its dynamic evolution with time.The results show that,firstly,in the emotional classification research,the sentiment classification of the convolutional neural network model is better than that of the support vector machine.After the parameter optimization,the classification accuracy of the model is significantly improved in the three emotional review texts.It can be seen that the convolutional neural network model is more suitable for emotional classification of Weibo text.Secondly,in the topic hotspot analysis based on sentiment classification,price,service and security are the main concerns of users on the network car.The network car platform has different concerns in different development stages.After the US group took a taxi to enter the Shanghai market,the user's concern was first to welcome the price subsidy,to support the US group to take a taxi into the market,to oppose the monopoly,and then the price subsidy disappeared,the network car was gradually exposed,the deck car,the female murder event and navigation Events such as detours occur frequently,and users turn their attention to safety and service.Based on the above conclusions,it is recommended that the network car platform should improve the procedures for entering the market,form a benign competition,strengthen industry standardization,and improve service quality.Consumers should improve their self-security awareness and supervise the development of the network car platform.