EVOLVING USER BEHAVIOUR PROFILES RECURSIVELY USING STATISTICAL COSINE DISTANCE

Abstract

Recognizing the behaviour of others in real time is a significant aspect of different human tasks in many different environments. When this process is carried out by software agents or robots, it is known as user modeling. The recognition of users can be very beneficial for assisting them or predicting their future actions. Most existing techniques for user recognition assume the availability of handcrafted user profiles, which encode the a-priori known behavioural repertoire of the observed user. However, the construction of effective user profiles is a difficult problem for different reasons: human behaviour is often erratic, and sometimes humans behave differently because of a change in their goals. This last problem makes necessary that the user profiles we create evolve.