Research And Application Of Car Reviews Text Classification Based On LSTM

With the rapid development of the Internet,social media and online shopping have become popular.At the same time,a large number of online review texts related to automobiles have appeared on the Internet.These car reviews contain a lot of information that is valuable to the company.Using appropriate classification techniques to classify them can bring great help to the company.However,there is relatively little research on the classification and application of online car review texts,especially the lack of adopting new technologies such as deep learning to explore the classification and application of online car review texts.For this reason,this paper aims at the problems existing in online car review text classification.It has carried out research on the classification and application of auto review texts based on LSTM,and has mainly completed the following work:(1)Constructed a car review text classification and application framework.Aiming at the online auto review text,it analyzes the impact on the production and sales of auto companies,sums up the related applications based on online auto review text mining,and focuses on the core issue of online auto review text classification,constructed a car review text classification and application framework based on deep learning.(2)Proposed car review text classification models based on LSTM.For the subject classification of car review texts,an LSTM model is proposed to deal with them.This model,with its memory cell structure,has an ideal effect in text sequence analysis tasks.And for the sentiment classification of car review texts,proposed a BLSTM model with sentiment word embedding as input.By constructing the semantic and sentiment dual input matrix,the emotion expression ability of the model is enhanced,and the accuracy of the online car review text sentiment classification is improved.(3)Designed and implemented a car review text mining system.On the basis of the proposed classification model,this paper proposes a text mining system for car review under the internet big data environment,analyzes the functional requirements and the overall architecture of the system,and designs the implementation flow of the main functional modules of the system in detail.Finally,the system’s visual interface is shown.