A Cooperation Based Metric for Mobile Applications Recommendation

As the smartphone is popular recently, mobile applications have been more and more attractive to users. With the exponential growth of applications' number, many works have been done to recommender the interesting applications for users. To check the effectiveness of these methods, the researchers are all tend to choose the traditional accuracy metrics to evaluate their system. However, considering the particularity of the mobile application recommendation, the accuracy evaluation metrics are not enough. Aiming at this problem, the goal of this paper is towards propose a new metric to evaluate the results of the mobile recommender systems and present an associated recommendation methodology. Base on the real usage pattern of mobile applications, we suggest a new evaluation metric called Cooperation. That is, when users use one application, they would enjoy using another one, we say those two applications have Cooperation Relation. In addition to, we also design and implement a strategy, which could improve the Cooperation degree of the mobile application recommendation results.