The Design And Implementation Of Recommendation System For Collaborative Crowdsourcing Testing

As an emerging software testing method,crowdsourcing testing can call a large number of workers and conduct online testing in a short time.Therefore it has received extensive attention in academic and industrial circles.Compared to traditional software testing,it can adapt to the testing needs of fast iterative software products in the mobile Internet era.At present,due to the lack of support from the crowdsourcing test platform,the research on crowdsourcing testing take action in model simulation or data analysis.These methods lack verification of the real scene.At this stage,crowd-sourcing test platforms adopt a competitive approach,resulting in a large number of duplicate and uneven reports.At the same time,due to the long tail effect,some testing requirements are not fully covered.In view of the problems existing in current crowdsourcing testing,this thesis proposes a recommendation system for collaborative crowdsourcing testing.During the entire testing process,users undertake both the test task and the audit task.The system performs real-time similar report recommendation when users filling in reports,guiding users to review the submitted report,thereby avoiding the submission of duplicate reports.After submitting reports,this system provides users with task assignment.The audit report recommendation provides users with a list of reports to be reviewed based on users'collaboration data and historical submission records.The test page recommendation visually suggests pages to be tested for users.For the recommendation report,users can perform likes or dislikes.This system uses cross-audit between users to verify the generality of problems in reports.At the same time,this system provides users with choices to modify the recommendation report,using multiple people editing together to improve the quality of the description in the report.This system is realized by mainstream frameworks,with Angular2 as the front-end framework,Spring Boot as the back-end framework,Redis as the query cache,and MongoDB as the database.Similar report recommendation is implemented by Word2Vec and WMD algorithm.The audit task recommendation is implemented using a model-based collaborative filtering approach.The test page recommendation is implemented using the multi-source shortest path method based on users' histories.This system is tested for function.Besides,it uses Jmeter to complete the stress test.The system is deployed and run on the crowdsourcing test platform.Competition with independent system showed that the system can effectively reduce the repeated reports.It also improves the quality of report text descriptions and reduces the time required for the full coverage of test pages.