Research On APP Software Defect Analysis Method Based On User Reviews

At present,with the development of the Internet,the Internet industry has been constantly concerned by everyone.The Internet industry,which is concerned,has also spawned a lot of APP software,which involves many aspects of life.With the rapid growth of the application scale of APP software,APP software is more and more closely related to people,and it will affect the work and life of the user because of the defects of the APP software itself.Different users may have different feelings in the process of using APP software,and users can publish the related feelings in the form of APP software user reviews in the application market.In these user reviews of APP software,some expressed their like for APP software while others expressed their disgust for APP software.In a large number of user reviews of APP software,some user reviews of APP software reflect the defects of APP software.The defect here reflects the poor experience of users in the process of using APP software,which deserves the attention of other users and developers.Therefore,how to analyze APP software defects based on user revi ews has become a problem that needs to be studied.Based on the above discussion,the following studies are carried out in this paper.(1)this paper first discusses the importance of APP software user reviews,then analyzes a large number of APP software user reviews,and classifies the APP software user reviews that embody the defects,summarizes the common APP software defect types and gives the corresponding examples of the APP software user reviews.(2)because user reviews of APP software containing d efects are likely to be included in negative comments,to improve the accuracy of APP software defect classification model,this paper uses emotional tendency analysis method to dig out negative comments from user reviews of APP software,and finally uses negative comments as a step(3)defect analysis data.(3)aiming at the negative reviews,we use the trained and tested defect classification models of APP software to analyze the defects.The defect classification model of APP software needs the process o f training and testing,in which the training process includes participle,disuse word processing,feature selection,and the model of APP software defect classification is built with the naive Bayes.The test process is mainly through ten fold cross valid ation to test the defect classification model APP software.