Research On Hybrid Recommendation Method For Third-Party Library Of Mobile Application Based On Knowledge Graph

With the popularity of smart terminal devices and the rapid development of mobile Internet,the mobile application market is increasingly prosperous,and the number of mobile applications and user groups is increasing rapidly.Although the rapid growth and continuous expansion of the software market have brought tremendous convenience to users,for the teams and developers who provide these software services,the challenges and competition they face are more intense.In order to meet the more demanding user requirements,developers need to develop more mobile applications that meet the user's functional requirements in a shorter period of time,while also ensuring the quality and usability of the applications.Third-party libraries play a vital role in the development of mobile applications.These libraries can greatly shorten development time,improve development efficiency and improve development quality.Choosing right third-party libraries can help developers complete mobile application development more efficiently.However,a large number of third-party libraries have been released so far,which puts a heavy burden on developers to choose the suitable third-party libraries.How to help developers make better choices is a question worthy of attention.For this problem,this paper proposes a novel hybrid recommendation method TM-MKR for third-party libraries of mobile applications based on knowledge graph.By combining knowledge graph technology and topic modeling technology,the interactive information,unstructured textual description information and structured semantic information of the applications and third-party libraries are used simultaneously to achieve hybrid recommendation.In order to verify the effectiveness of the method,we crawled data from AppBrain and GitHub and constructed two real-word datasets.In the process of building the GitHub dataset,in order to detect the third-party libraries used in the applications,we also study the existing third-party library detection methods,and proposes a third-party library detection method for the specific scenarios of the dataset.Based on these two datasets,two large-scale domain-specific knowledge graphs were constructed using graph database technology.Based on a series of experiments conducted on these datasets and knowledge graphs,through comparison with several state-of-the-art recommendation methods on multiple metrics,the experimental results show the effectiveness of the proposed method in the recommendation scenario of third-party libraries for mobile applications.