Towards Efficient Reality Mining with Contexts and Semantics: A Case Study of Telecommunication

Recommendation systems need to employ efficient reality mining module to find out whether social relationships extracted from a collected dataset are really meaningful in real world. Also, the recommendation should be generated by taking current contexts (e.g., location and time) account into. In this paper, we want to propose an ontology-based framework to discover real social networks from a large amount of dataset. This framework can provide a semiautomatic approach to build meaningful social networks by repeating interactions with human experts. We assume that given a dataset, we have to discover the hidden social networks which express the contextual dependencies between people. The personal context of a certain person is interrelated with those of other people, and investigated how to take into account his neighborpsilas contexts, which possibly have an important influence on his personal context. Particularly, in this research project, the target dataset in this work is collected from user activities (about over two millions of people) in telecommunication industry. We have applied the proposed system to discover the social networks among the mobile subscribers to provide recommendation to mobile devices.