Comment authenticity detection method and system

Abstract

The invention discloses a comment authenticity detection method. The method comprises the steps of performing word segmentation preprocessing on a to-be-detected comment to obtain multiple word segmentation results; converting words in the word segmentation results into word vectors; calculating a sentiment label of the to-be-detected comment through a pre-built convolutional neural network model,wherein the convolutional neural network model is generated by training a convolutional neural network through word vectors and sentiment labels of standard sentences of a training set; selecting a target eigenvector from all to-be-selected variables by adopting a logic regression model; calculating a distance between the target eigenvector of the to-be-detected comment and a clustering center ofa real comment, and calculating a distance between the target eigenvector of the to-be-detected comment and a clustering center of a false comment; and according to the distances and the sentiment label of the to-be-detected comment, determining the authenticity of the to-be-detected comment. The accuracy of detecting the authenticity of the comment can be improved. Meanwhile, the invention furthermore provides a comment authenticity detection system.

Description

technical field [0001] The invention relates to the technical field of e-commerce, in particular to a comment authenticity detection method and system. Background technique [0002] With the advent and vigorous development of the web2.0 era, users are increasingly using various e-commerce platforms for shopping and opinion sharing. [0003] However, most e-commerce platforms have major flaws. These platforms allow users from all over the country or even the world to make unlimited comments, which has prompted some unscrupulous merchants or manufacturers to hire some brushers to post some positive comments to influence the judgment of potential consumers and increase their sales. quantity, or publish some negative comments in the competitor's online store to affect the reputation of the competitor. Since the structure of these fake reviews is often very similar to real reviews, it is difficult for consumers to identify these harmful fake reviews when reading these reviews. ...

AI Technical Summary

Problems solved by technology

However, due to the noise of user comments, many new words, their own fixed collocations, and limited context information, it is very difficult to analyze the tendency of user comments.

Images