Design And Implementation Of A Mobile Device Combination Recommendation System For Mobile Application Compatibility Testing

With the frequent updates of mobile devices,the increasing number of mobile devices in the mobile device market and the serious differences in software and hardware require mobile applications to conduct compatibility tests on a larger and more complex combination of mobile devices,which is a severe challenge for mobile application compatibility testing.Aiming at the problem of mobile device combination selection during mobile application compatibility testing,this thesis is dedicated to recommending to testers a combination of mobile devices that have a high degree of matching to the market and a small-scale compatibility combination.Based on the knowledge-based recommendation technology,this thesis designs and implements a mobile device combination recommendation system for mobile application compatibility testing.The main work of this thesis includes:1)Design and implement recommendation system.This thesis analyzes the requirements of the actual process of compatibility testing,and designs the overall system,finally designs and realizes a mobile device combination recommendation system for mobile application compatibility testing with the recommendation module and the expert knowledge base module as the core modules.Among them,the recommendation module is the core business logic module of the system,providing users with compatibility testing mobile device combination recommendation functions;the expert knowledge base module generates expert knowledge and provides the knowledge for the recommendation module.2)The detailed design and implementation of the recommended module.On the one hand,according to the knowledge-based recommendation technology,this thesis designs the recommendation model and recommendation algorithm in detail,and compares the model's mobile device combination recommendation results for compatibility test with the recommendation results of human experts.The experimental results show that the recommendation model can recommend testers a mobile device combination to do compatibility testing effectively.On the other hand,this thesis designs and implements three sub-modules in detail based on the recommendation model: recommendation sub-module,recommendation report sub-module and recommendation utils sub-module.The recommendation sub-module is the core part,which generates recommendation results based on the expert knowledge of the expert knowledge base module and the recommendation model;the recommendation report sub-module generates recommendation reports based on the recommendation results and provides users with the recommendation report query function;the recommendation utils sub-module offers auxiliary calculation functions to other two sub-module.3)The detailed design and implementation of the expert knowledge base module.Through interviews with test experts,this thesis summarizes the expert experience,and designs the knowledge representation,knowledge storage and knowledge inferring of expert knowledge in detail,and finally realizes the ability to infer mobile device combination for compatibility test according to mobile device market distribution data.This module offers the knowledge to other modules with query interface.The mobile device combination recommendation system for mobile application compatibility testing designed and implemented in this thesis has been deployed on the internal server and has passed functional and non-functional testing.The system has reached a recommendation accuracy rate of more than 90% in terms of recommendation effects and can help testers choose a combination of compatible test mobile devices that satisfy expert standards.