The increasing flourish of available services in telecom domain offers more choices to the end user. On the other hand, such wide offer cannot be completely evaluated by the user, and some services may pass unobserved even if useful. To face this issue, the usage of recommendation systems in telecom domain is growing, to directly notify the user about the presence of services which may meet user interests. Recommendation can be seen as an advanced form of personalization, because user preferences are used to predict the interests of users for a new service. In this paper we propose a recommender system for users of telecom services, based on different collaborative filtering algorithms applied to a complex data-set of telecom users. Experiments on the recommendation performance and accuracy are conducted to test the different effects of different algorithms on a data set coming from the telecom domain.
展开▼
