Research On Copycat Application Detection Based On Text Clustering Analysis

With the development of mobile information technology,the number of users on the mobile application platform is increasing day by day,and the mobile application market is showing an explosive growth trend.This provides huge market opportunities and economic motivation for mobile application developers to develop novel applications.But at the same time,a problem inevitably arises — "Copycat" applications.According to the "Android mobile app copycatting Research Report in2015" released by 360,mobile app copycatting is very rampant,with an average of 93 copycat apps corresponding to each legitimate app.People resist the emergence of "Copycat" applications.The names or application descriptions of these applications are very confusing and easily misunderstood by consumers.For example,the popular mobile game "Fortnite",which is popular with players because of its novel escape genre,has been downloaded millions of times and topped the annual mobile game charts.But at the same time as the game’s popularity,more than 100 similar game applications have entered the Google Play store,causing distress for developers of original apps.They called on the platform to take action against "Copycat" applications,which have stolen their innovation achievements and infringed on their economic interests as well as hit their innovation motivation to a certain extent.It’s much more difficult than it seems to solve the problems caused by "Copycat" applications.The main reason for this situation is the lack of an objective method to identify whether an application is an original application or a "Copycat" application.In order to solve this problem,this paper proposes a practical method to identify "Copycat" applications by combining machine learning techniques such as natural language processing,ensemble learning,topic modeling,spectral clustering,and affinity propagation algorithm.In addition,the empirical study also found that "Copycat" applications can promote the performance of original application,and the quality of "Copycat" applications can enhance its substitution effect on original applications.This paper contributes to the research of "Copycat" applications in the application market in the Chinese context.