Bayesian Filters for Mobile Recommender Systems

With millions of registered users and a phenomenal increase in the amount of information exchange between them, it is clear that social networks are becoming increasingly popular on the Internet. Due to this popularity, they have become an important platform for business and marketing activities. In this paper, we understand the availability and the need of social networking sites with the mobile environment. The existing mobile recommender systems focus on targeting the users based on predicted demographics rather than the interests of the user. By extracting users' information from social networking sites, we design four novel filters to improve the targeting accuracy of the mobile service providers. Finally we have applied the Bayesian approach on the filters to arrive at a proper response percentage of the targeted customers. In essence, we ensure that based on applied strategy of understanding the customer's unique preferences, mobile service providers can launch specific campaigns on selected groups of users to get useful feedback. Moreover, it improves customer loyalty towards the operator. We have evaluated our findings through standard techniques on real world data.