Title:
|
AN APPROACH TO BUILDING PROFILE BASED ON THE UTILITY OF FEATURES IN THE ADAPTIVE SYSTEMS |
Author(s):
|
Thi-huyen Le , Duc-long Le , An-te Nguyen , Dinh-thuc Nguyen |
ISBN:
|
978-972-8924-93-5 |
Editors:
|
Pedro Isaías, Bebo White and Miguel Baptista Nunes |
Year:
|
2009 |
Edition:
|
1 |
Keywords:
|
Adaptive system, user profile, useless feature, useful feature, extra value, rare value |
Type:
|
Full Paper |
First Page:
|
295 |
Last Page:
|
302 |
Language:
|
English |
Cover:
|
|
Full Contents:
|
click to dowload
|
Paper Abstract:
|
Recently, adaptive systems have been employed in many fields such as e-Commerce, recommender system and e-
Learning. In these systems, every user owns a profile to represent his specific individual characteristics. Depending on
the different applying field, it contains different information to describe user such as personal identification, backgrounds,
demands, preferences, habits and so on. Through that information, the system provides for user with suitable resources
and services. Therefore, the quality of recommendations mainly depends on user profiles and many researches in study
have proposed the generic model of user profile for different applying fields. But developers still meet many difficulties
to customize the generic model for a specific system of each applying field because profile can be either redundant if it
includes an amount of features more than features which the system can be able to provide, or deficient if it has an
amount of features less than features which profile can represent completely an user.
In the adaptive system, profile is used to represent individual characteristics of user. So clearly, if a feature of profile,
which its value is all the same with every user, then it could be called that useless feature. On the contrary, a feature
could be considered as useful feature if it is able to distinguish between users. . Our paper presents an approach to
building profile based on the utility of features. This approach includes two phases: determining and rejecting the useless
features, then dealing with the values of feature that one is considered as an unwantedvalue from useful features. |
|
|
|
|