We are developing a system which assists users by collaboration between the users and environment. Our collaboration system provides services according to user behavior proactively in homes when environment detects high-level user behavior such as "leaving the home". To realize such a collaboration system, this paper proposes a method for detecting high-level user behavior. The proposed method dynamically sets values suitable for individual behavioral pattern of each user to thresholds used for detection. A conventional method determines threshold values common to all users. However, the common values are not always suitable for all users. Our method determines threshold values suitable for a user by utilizing data of other users whose characteristics are similar to the user, with collaborative filtering. © 2008 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Yamahara, H., Harada, F., Takada, H., & Shimakawa, H. (2008). Personalizing threshold values on behavior detection with collaborative filtering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5061 LNCS, pp. 411–425). https://doi.org/10.1007/978-3-540-69293-5_33
Mendeley helps you to discover research relevant for your work.