Research On Sentiment Analysis Of The Shopping Reviews Based On The E-commercial Platform

In recent years,the Internet has led to the rise of online shopping malls.More users choose to buy products online and shopping reviews have gradually become the main channels of information dissemination.Users often make decisions based on the comments and leave comments to share information.Hence,the sentiment analysis of online shopping reviews has become an important research topic.The sentiment analysis of shopping reviews can help businesses and users accurately understand the advantages and disadvantages of products and provide them with strong references.However,due to the large amount of reviews,it is difficult for users to extract useful information.Therefore,this paper adopts the method of using sentiment dictionary based on semantic knowledge to analyze the emotional tendency of shopping reviews.After discussion and research over the issue,the experimental results have proved that the proposed method in this paper can improve the accuracy of emotional analysis.Our completed work includes the following:(1)This paper analyzes the characteristics of several key technologies in the related field comparably.And summarizes the advantages and disadvantages of word segmentation and feature extraction methods frequently used in the present research.The research theory,mathematical expression and its characteristics are briefly introduced for following research.(2)This paper preprocesses the reviews and removes the stop words to reduce the impact of noises during the sentiment classification.Considering that the phrases used in the shopping reviews of various fields may differ from each other,the proposed PTF-IDF method is to construct the domain emotional dictionary to obtain more emotional characteristics.(3)Emotional elements such as network language,emoticons are comprehensively considered.The degree words,negative words and double negative words in the Chinese syntax that used to modify the reviews are handled appropriately.In order to quantitate the relationship between the emotional words and the domain reviews,the proposed Field-PMI method is to measure their relevance and make further analysis on the achieved domain emotional words.Theoretical analysis and extensive experimental results show that the proposed methods can efficiently improve the emotional classification accuracy of the shopping reviews.