Comprehensive Evaluation Of E-commerce Products Based On Multi-dimensional Information Fusion And Its Applications

With the continuous development of Internet technology,e-commerce has been used widely,and e-commerce platform has become an important channel for online shopping.However,the e-commerce platform has also exposed some problems,such as counterfeiting,false propaganda,and scrutiny,these infringe on consumer rights and undermine normal business order.Considering these problems of ecommerce platform,how to correctly judge the quality status of the goods in the ecommerce platform according to the existing data information given at the ecommerce platform,and guide the consumers to purchase correctly are particularly important.Based on the data & information of multiple aspects of products at the ecommerce platform,this thesis proposes a comprehensive scoring model of ecommerce products based on multi-dimensional information fusion.The main research contents of this thesis are as follows:(1)Calculation of merchant reputation and product popularity.This paper analyzes the business information in the e-commerce platform,and summarizes the business reputation formula;the commodity-related information is numerous,including monthly sales,collections,comments,prices,etc.select the hotspot information that consumers pay attention to,and calculate the product popularity.(2)Extracting and clustering of product features.Based on the lexical features,domain features and dictionary features of feature words,this thesis uses CRF method to identify and extract feature words.Then the traditional semantic dictionary similarity is combined with the word vector similarity,and the K-Means++ algorithm is used to cluster the product feature words.(3)Emoting analysis of product characteristics.Integrate existing online sentiment dictionaries and quantify the emotional intensity of sentiment dictionaries.Based on the seven-tuple(F,N,A,C,E,L,W)descriptions proposed in this thesis,the emotional scores are calculated for product features,where F is the product feature word,N is the negative word,and A is the degree adverb.C is a conjunction,E is an emotional word(comment word),L is the clustering category corresponding to the product feature word,and W is the product feature weight,corresponding to the TFIDF value of each feature word.(4)Multi-dimensional information fusion model and system development of ecommerce products.The multi-dimensional information evaluation score of the ecommerce product is integrated,and the product comprehensive evaluation score is finally obtained.Based on the research theory of this thesis,a decision-making system based on multi-dimensional information fusion of e-commerce products is designed and implemented.This thesis combines the business reputation,product information,consumer evaluation and other dimensions to comprehensively evaluate the products sold on the e-commerce platform,truly reflecting the quality and safety of the e-commerce products,so that consumers can purchase products.Provide a reference basis to reduce the risks in the online shopping process.