Research And Implement Of APP Software Evaluation Method Based On User Reviews

With the rapid development of the society,the advance of mobile intelligent terminal has gradually exceeded the personal computer,while the growth of the APP software erupts.The APP stores came out,and the difficulty that finding the APP software has been eliminated.However,it is not comprehensive for APP stores to audit the APP software.And inferior,counterfeit,malicious APP software takes advantage of a weak point which bringing unnecessary troubles to the numerous users.Therefore,it is so significant to study how to judge the APP software.And users evaluate the APP software on making reviews,which leads to analysis whether the APP software is good through the users' reviews.But there is no restriction on the user reviews of APP software in the APP stores,which is also an obstacle to user reviews.Therefore,it is of great signification to study the APP software evaluation.APP software is different from general software.Because of the imperfection of APP software evaluation methods on users,it has the following problems:(1)APP software has the characteristics of which general software does not have,while evaluating the APP software cannot use the method of the general software evaluation at all.(2)The user reviews of APP software are not only short,high-generality,but also informal,free which makes evaluating the APP software and browsing the user reviews difficult.In order to solve the above problems,this paper put forward to filter APP software user reviews and extract the words which describe the features of APP software or users'affecting,and the APP software evaluation method based on the words.At last,evaluate the APP software by using the filtered reviews.During this period,the following work has been completed:(1)Because of the freedom and arbitrariness of the APP software user reviews,a software user reviews relevance judgment method is proposed based on Naive Bayesian classification algorithm and term statistical method APP.And the feature-lexicons of APP software are obtained;(2)An evaluation model for APP software is proposed.(3)The metric method of APP software model indices is founded.(4)The weights are evaluated by the grey relational analysis of the grey system theory.(5)The effectiveness of the methods is proved and verified.