Learning Algorithm for Human Activity Detection in Smart Environments

Anticipating users´ actions and automating them is an important goal in intelligent environments. This work presents IntelliDomo´s learning algorithm, a data mining approach applied to home automation systems. This algorithm is able to acquire users´ habits just by processing the data collected from the full environment. IntelliDomo can generate production rules for each behavior pattern in order to anticipate the user´s periodic actions. This paper presents the fundamentals of this algorithm together with the description of the main parameters used.