A Two-stage Personnel Recommendation Method Directed Towards Mobile Crowdsourcing Testing

Abstract

Nowadays, millions of mobile applications bring convenience to people’s lives, but at the same time, real scenes are difficult to simulate, recruitment of testers and other problems. At present, numerous mobile application crowdsourcing test platforms are emerging. They effectively solve the problem of mobile application testing by recruiting a large number of testers online and using real scene testing. However, with the registration of a large number of testers and the release of test tasks, the platform is faced with the problem that task publishers have difficulty in selecting high-quality personnel and personalized matching tasks, which severely restricts the development of mobile application crowdsourcing testing. Aiming at these problems, this paper firstly analyzes the characteristics of mobile application crowdsourcing test. Then the classifier indexer, personnel predictor and task predictor are generated. On this basis, two-stage staff recommendation algorithm tstage-pi and two-stage task recommendation algorithm tstage-ti are proposed to solve these problems. Through the two algorithms proposed in this paper, the matching degree between testers and crowdsourcing testing tasks can be greatly improved.