Recognition Of Fake Online Products' Reviews

With the rapid development of e-commerce,consumers are keen on online shopping and publishing reviews,these reviews provide important reference information for stores,manufacturers and potential consumers.However,the reviews may be false,and the existence of false reviews will influence the reference value to a certain extent.Therefore,it is extremely important to identify false e-commerce reviews.This article mainly focuses on the e-commerce reviews,focusing on the methods of reviewing explicit features,implicit features,and the combinations of both.The following research work has been completed:(1)The current false reviews recognition is performed through explicit features of reviewer or review text,but does not consider the semantic information of review text,and the analysis result is not accurate enough.Therefore,an electronic business false review recognition method based on implicit semantic analysis is proposed.Based on the analysis of user behavior characteristics,this method adds semantic analysis information of review text,making it possible to identify false reviews at the semantic level of the text and improve the recognition accuracy.(2)Analyze review text using dependency syntax,extract the attribute value pairs in reviews,and then further filter the feature value pairs with more reference value through the three methods of dependency relations,part-of-speech rules and dictionary.The number of feature value pairs is an important reference item for identifying false reviews.(3)Use the dislocation level method to evaluate the experimental results.Due to the particularity of experimental method,it is not convenient to use traditional assessment methods.Therefore,thedislocation-level assessment method was used in this paper.(4)Finally,an e-commerce review analysis system was designed and implemented,which facilitates further analysis of reviews and research methods for identifying false reviews.