Design and implementation of user context aware recommendation engine for mobile using Bayesian network, fuzzy logic and rule base

The purpose of this paper is to propose an intelligent service recommendation model. The paper formulates the service adaptation process by using artificial intelligence techniques like Bayesian Network, fuzzy logic and rule based reasoning.

The authors formulate the service adaptation process by using artificial intelligence techniques like Bayesian Network, fuzzy logic and rule based reasoning. Bayesian Network is used to classify the incoming call (high priority call, low priority call and unknown calls), fuzzy linguistic variables and membership degrees to define the context situations, the rules for adopting the policies of implementing a service, fitness degree computation and service recommendation. In addition to this the paper proposes maximum to minimum priority based context attributes matching algorithm for rule selection based on fitness degree of rules. The context aware mobile is tested for library and class room scenario to exemplify the proposed service recommendation engine and demonstrate its effectiveness.

First, it was found that there was reduction in application searching time in different contexts. For example, if user enters into the library, the proposed mobile will be adapted to the library situation automatically by configuring its desktop and internal settings to facilitate the library services like book search, web link, silent mode and friends search. Second, the design of the recommendation engine, utilizing contextual parameters like Location (class room, college campus, house, etc.) Personal (age, name), Temporal (time, date), Physical (fall, normal), and schedule agendas, was found to be of importance.

Exploitation of hybrid fuzzy system, Bayesian Networks and the utility theory (usage history and context history) for modeling and implementation.