The Identification Of Electronic Commerce False Comments Based On DBN Model

In recent years,with the rapid development of China's e-commerce industry,people's dependence on online shopping is also increasing.Young consumers' habits and understanding of consumption have long been no longer simply offline shopping.Today,online shopping has become the preferred method for most young people to shop.They enjoy this convenient shopping experience.However,there are still some serious problems in the development of the e-commerce market in China at present.For example,due to the characteristics of the e-commerce market,its entry threshold is low,a large number of shops are mixed,and transaction disputes caused by the asymmetric information between the store and the consumer also occur frequently.The time and space isolation of the transaction process leads to incomplete information acquisition;the isolation of logistics and business flow leads to the entire transaction involving multiple industries to facilitate fraud.At present,corresponding to the booming development of China's e-commerce is the matching network transaction integrity system has not been fully established and improved.This has led to a number of unscrupulous merchants using this feature of the e-commerce market to pay large amounts of money,thereby increasing the behavior of their own shop credits and sales of their own goods.These false information have seriously affected people's buying decisions and brought extremely bad influence to society and the market.False trading,especially the recognition of false positive feedback,is beneficial to consumers' individual self-recognition ability and cost saving;it is conducive to maintaining and standardizing the healthy development of the e-commerce industry.This article analyzes the whole process and characteristics of merchants' billing from production to implementation on the basis of sorting out and sorting out the research results of a series of online trade false recognitions and recognitions at home and abroad.Based on these characteristics,a large number of product reviews were obtained using the distributed crawler technology,and a comparison was made on the information that was partially grasped in advance in the hands.Secondly,the data on the data was collected from a large number of commodity reviews and the data set was annotated;Taking the feature set of product reviews as an entry point,the deep confidence network algorithm based on deep learning is used to analyze and identify the credibility of the e-commerce transaction result,that is,the product reviews;finally,the accuracy of the model is verified and compared with other shallow machine learning algorithms.The accuracy rate of therecognition of the deep confidence network for the review data is significantly higher than that of other shallow machine learning algorithms.Based on the results of the model recognition,the review characteristics of normal consumers are analyzed.The main contributions of this paper are as follows:(1)Using the product review data as a breakthrough point to explore the language features of consumer product reviews.Taking the review data as a breakthrough point,it can effectively portray the language features of consumer review data,and expect to discover the language features of normal consumers when they review the product.(2)Based on the deep confidence network algorithm,it can effectively identify e-commerce false comments.In the past,many studies used computer to simulate the occurrence of false comments.This article uses a Python-based crawler technology to obtain massive real user review data,and uses the deep confidence network algorithm to identify the single review.