Sentiment Analysis Research And Application Of Online Shopping Review Data

With the further spread of the Internet,more and more people are beginning to contact the Internet,and more and more people are starting to shop online.When people shop online,they often comment on the products they buy.Therefore,there are a large number of consumer product reviews on various products on the e-commerce website.Commodity reviews contain real opinions of consumers about the products they buy.Through these comments,we can know how other consumers like a certain product,and also can compare a certain category of goods horizontally to know The advantages and disadvantages of each product.So deep exploration of the content of these reviews can get a lot of useful help for life.Therefore,this article takes the mobile phone commodity review data as an example to conduct emotional analysis on the review data.Then the emotional score calculation is performed on the results of the sentiment analysis,so that the consumer can quickly and easily understand the evaluation of the other consumption of the product,thereby understanding the main advantages and disadvantages of the product and helping them to make purchase choices.This article uses Python to write the web crawler program,and collects the comment texts of some mainstream mobile phones and the basic information of the reviewers as the original data in Jingdong.By combining the nickname of the reviewer,the user level,the time of placing the order,and the time of the comment,the original data is deduplicated and deactivated.Then,the Chinese word segmentation of the processed data and the construction of the LDA theme selection model are combined with the actual situation to extract the five attributes that consumers pay most attention to.By constructing the Word2 vec word vector training model,the text data is extracted,and the text data is transformed into a vector form that can be directly input into the sentiment analysis model.Then,the paper analyzes the five different attributes of the comment data and constructs and trains the LSTM sentiment analysis model to analyze the emotions of the mobile phone comment data.In the end,this paper analyzes the sentiment scores of five mobile phone attributes with the similar price of Xiaomi 8,Glory 10 and Huawei P20,which are slightly higher than their Huawei P20.Analysis allows consumers to quickly understand the pros and cons of each phone and help them make purchasing decisions.