Context prediction has been receiving considerable attention in the last years. This research area seems to be the next logical step in context-aware computing, which, until a few years ago, had been concerned more with the present and the past temporal dimensions. Most of research works related to context prediction employ the same algorithm for all cases. We did not find any approach that automatically decides the best prediction method according to the situation. Therefore, we propose the ORACON model. ORACON adapts itself in order to apply the best algorithm to the case. This adaptive behavior is the main contribution of this work and differentiates the proposed model of other related works. Furthermore, ORACON supports other important aspects of ubiquitous computing, such as, context formal representation and privacy. We have built a functional prototype that allowed us to conduct two experiments. The first experiment successfully tested the main functionalities provided by ORACON to support context prediction and privacy aspects. The test used context histories generated with a location database that contains 22 millions chekins across 220,000 users in the location sharing services Foursquare and Twitter. The second experiment assessed the adaptive feature of the ORACON. The test simulated the behavior of 30 users for a period of 30 days, using context histories generated through the Siafu simulator. This tool generates data for the evaluation and the comparison of machine learning methods in mobile context-aware settings. We concluded that ORACON chose the most accurate prediction algorithm in the simulated scenario, proving that the model reached the main contribution sought by this research.

Expert Systems with Applications