Abstract: The technology of mobile cloud computing is benefit for deploying various mobile applications. However, there is an energy consumption problem to access cloud resources via mobile phone, which needs to establish connections many times under unstable communication conditions. To solve this problem, a link prediction method based on maximum user interaction behavior was proposed. Firstly, based on data prediction model, an interaction degree method based on improved interaction relationship is used to predict the data accessed by users. Then, combined with the friend link method of social network based on user behavior, the prediction data is analyzed and filtered, and the pre-storage mechanism is used to pre-store the above prediction data. Experiments show that the expected energy saving of mobile phones can be achieved without involving users' private information and improving the hit rate of users' next visit.

Key words: smartphones save energy, access data prediction, social network, interaction times