Improving Activity Mining in a Smart Home using Uncertain and Temporal Databases

Josky Aïzan, Cina Motamed, Eugene Ezin


In the context of smart home, activity mining appears as an interesting and promising solution for learning activity of daily living. This paper is an extension of a previous a research work titled Activity Mining in a Smart Home from Sequential and Temporal Databases. It proposes an activity mining method based on uncertain and temporal sequential pattern mining to deal with data uncertainty and events temporal relationships. It allows to track regular activities and to detect changes in an individual’s behavioural pattern. Uncertain sequential pattern mining algorithm is firstly applied to the input sequence database to extract typical sequences and secondly a clustering approach based on sequence alignment methods is performed in order to obtain separated typical activities. The results obtained are enough good compared to existing related works.


Paper Citation