Fuzzy method to disclose behaviour patterns in a Tagged World

The main objective of the Ubiquitous computing is to include technology on the user life without modifying their daily routine. A lot of kinds of applications base their operations on using sensors situated on a Tagged World, which is a smart area that serves to recognize user’s behaviour using information about their daily activity. It manages information collected by sensors to identify the user behaviour and to provide some services according to inferred behaviours. In this paper, we present a specific method to extract behaviour patterns on time from the collected sensor signals. The method is based on frequent itemsets that represent the common actions of the user. After, we obtain sequence patterns from these extracted itemsets. Our method should pay attention to the user activities because these have non-random imprecision by definition. Thus, we have designed a method to handle this imprecision, through establishing a temporal constraint that is called Fuzzy Temporal Window.