J. Yu and M. Jeon (Korea)
Context-aware applications, AntMiner classification, userprofile learning, recommender system, ubiquitous computing.
With the promotion of ubiquitous computing environment,
it has become important for recommender systems to take
advantage of contextual information to provide preferable
items to a user. According to the current user-centric
context, such as a user’s degree of stress, changes in
ubiquitous environment, the user preference of items also
changes. However, traditional recommender systems just
offer preferable items based on the history of user’s
service patterns and user preference profile. In this paper,
we propose a context-aware intelligent recommender
system to provide preferable media service and contents
to a user using the user-centric contexts such as 5W1H
(Who, When, Where, Why, What, How). The proposed
recommender is composed of two detailed components.
One is the media service recommender using the history
of user’s service patterns and current user-centric contexts.
The other is the media contents recommender which is
based on the content-based information filtering using
contextual information. This paper shows the efficiency
of the media service recommender with AntMiner
classifier and the media contents recommender with the
content-based information filtering approach.