Measurement of Situation Similarity

Context-aware computing applications often perform an assessment of similarity between the user’s current situation and the predefined or learned situations in order to trigger appropriate adaptation rules. Because future situations are often not exactly matching with past ones (e.g., time continuously changes), the assessment of similarity between situations is not a straightforward task. This paper reviews approaches to this problem and investigates various measures for similarity in the context of situation similarity. As a result, we propose a novel measurement of situation similarity. The main different between our approach from the existing ones is that we take into consideration of context in which the comparison of two situations occurs. The evaluation shows that our method achieves good results in comparison with others.