On Text Sentiment Classification Based On CRNN-Attention Model

With the rapid development of information technology and the Internet,people are increasingly inclined to shop on the Internet e-commerce platform.The product review information filled out by the consumer e-commerce platform shopping reflects the consumer's emotional attitude towards all aspects of the product,such as whether the performance of the mobile phone product can meet the needs of the consumer and whether the pricing is reasonable.Text sentiment classification is an important research direction in the field of NLP.In recent years,many domestic and foreign scholars have carried out a lot of research around this topic.Applying text sentiment analysis to mobile phone product reviews can reveal people's emotional attitudes towards various attributes of mobile phones,and provide accurate reference for consumers to shop and improve products.Therefore,it is practical to study the comments of customers who purchase mobile phones.significance.This paper focuses on the relevant theories in the field of text sentiment analysis,applying some of the best models in the field and applying other models to compare and analyze,and then derive the direction of model improvement.Combining the text sentiment classification model with the LDA topic model provides a path for how to dig deeper into the data sets generated in the actual production process.The full text is divided into four parts: the first part introduces the meaning of the topic and the literature review at home and abroad.The second part introduces the basis of text sentiment classification,namely text segmentation,word vectorization,and introduces TextCNN,RNN-Attention and CRNN-Attention,three neural network-based text sentiment classification models;the third part preprocesses the reviews_Electronics dataset of Amazon reviews dataset and the crawled Huawei P20 mobile phone comment dataset to construct a suitable corpus,using TF-IDF Adjust the weight of the words and select the words with larger weights for feature extraction.Applying the classification model CRNN-Attention in the speech domain to the field of text sentiment classification,comparing the classification effects of the three models of TextCNN,RNN-Attention and CRNN-Attention in the above two processed data sets;the fourth part,Combining the text sentiment classification model CRNN-Attention with the LDA theme model,the emotional tendency of each topic of Huawei P20 mobile phone comment data set is deeply explored,and provides guidance for Huawei P series mobile phone update.