Intelligent configuration recommendation of context-aware mobile application

Automatic contextual reconfiguration, which is the process of application to adjust features due to context changes, is vital to mobile computing. Existing methods have common deficiencies: (1) expert knowledge used in model has deviation with real data, (2) can not deal with the fuzziness of context, (3) neglecting the personal frequent reconfiguration patterns. Hence, we are motivated to propose an intelligent configuration recommendation method of context-aware mobile application (ICR). ICR uses fuzzy rough set to acquire the relation between context and configuration from records, and then recommends configuration, which takes the configuration fitness, components consistency and user´s personal reconfiguring custom into account. We experiments with real data set, the results show that ICR can acquire fuzzy information contained in training set and recommend configurations, which is not only proper for current context, but also flexible enough for possible reconfigurations.