Ranking Smartphone Apps Based on Users’ Behavior Records
- 561 Downloads
This paper discusses the ranking problem of how to assign a linear order to a given collection of smartphone apps by using users’ behavior records. The desired linear order is thought of as an observable representation of users’ latent aggregate preference of the apps, which balances their conflicting individual preferences as far as possible. The app ranking problem can be interpreted as a stochastic acyclic subgraph problem on a complete digraph with each arc associated with a constant weight obtained from a known probability distribution.
KeywordsUsage pattern Users’ preference Bayesian inference Stochastic acyclic subgraph problem
- 1.Fujimoto M, Luo S, Sari MP, Shigeno M, Zago M (2014) Usage pattern analysis for Android applications: classification and measurement. In: The 2th domestic conference on serviceology, Hakodate, Hokkaidou, Japan, 28–29 AprilGoogle Scholar