Methods On Risk Level Assessment Of Online Shopping Based On Improved Bayes Algorithm And Its Applications

Along with the rapid development of e-commerce,some problems such as damaging the benefit of consumers are always happened,and cause great losses.These attracted the attention of the government,enterprises and individuals.During online shopping process,consumers often read and analyze the reviews to find the existing risks,and make the best choice.It will improve online shopping experience in the end.For these reasons,this thesis,using big data technology,proposes an approach of evaluating the risk of online shopping,and develops correspondence system.The main contents of this thesis are as following:(1)Emotional analysis of comments.In this thesis,a method of emotion analysis of reviews based on the combination of emotional viewpoint dictionary and CRF model is proposed.The CRF model is used to extract emotional opinion words from the processed reviews,and through manual screening,it is used to expand the existing emotional dictionary and mark the intensity of emotional opinion words.Establish the evaluation score calculation model,and verify the validity.(2)Clustering of commodity data.According to the analysis of e-commerce data,the K-means algorithm is set up with four basis points by method on median of average partition,and reset the center points by class median,that reduces the influence of outliers and achieves better clustering effect.(3)Establishment of risk assessment model for online shopping.The difference of index importance is analyzed and the influence of prior probability difference on the results of weighted Bayesian algorithm is solved emphatically.A new weighting method is proposed,and contrastively verify the improved weighted Bayesian algorithm has higher evaluation accuracy.(4)The development of online shopping risk assessment system.Based on the research results of this thesis,a system is developed to realize the application of risk assessment of online shopping and provides the decision-making evidences for customers during their online shopping process.This thesis excavates useful information from e-commerce platform,and gives consumers the risk level of online shopping through data analysis.It not only saves the time of comparison,but also improves the comprehensiveness of evaluation,so as to avoid purchasing higher-risk goods,thus,by restraining sales,crack down on bad businesses and improve the environment of online shopping.