Smart home user identification using bag of events approach

In this paper we present an innovative approach to identify people in smart homes. The behavioral pattern of smart home's users can be extracted to distinguish each user from others. In the lowest level, a sequence of sensor events forms an activity. In the higher levels, a combination of activities creates a behavioral pattern. The bag of events can be considered as features. In other hand, the behavior is flexible and can vary in different situations. A bag of activity approach is introduced to overcome the problem of behavior inconsistency. Our experiments confirm that the bag of activity is increased the f-measure of the model to near 98%.