The prediction of the context and especially the location of a user is a very important task in the field of pervasive computing. It is an important factor that can reveal the user’s needs and hence allows the proactive adaptation of services. This paper proposes a new approach to predict the future location in a pervasive system. It is a reasoning approach based on context-aware using an ontological probabilistic model. The process is performed with the use of an ontological model containing different applicable scenarios, contexts, the Bayesian network models, and describes the environment where a pervasive system exists. Both ontology and Bayesian network will be used to handle a stochastic process for prediction.
Tested on real data, our model was able to achieve 90% of the future locations prediction accurately. The model can be applicable in different use cases on smart space such as smart house, smart building or smart city. It will be useful in many fields for instance energy saver, education, mobility, and assistance for people with disabilities. In a future work, a real system (smart home) will be developed using the approach presented in this paper to assist children with autism.