Title
Real Time Learning of Behaviour Features for Personalised Interest Assessment
Abstract
This paper deals with an adaptive and personalized algorithm to dynamically determine the interest of a user in a document from the observation of his behaviour during its consultation. Several existing works propose an implicit feedback of user's interest from observations of his behaviours when he is reading a document. Nevertheless, the used algorithms are a priori defined and then cannot be really adapted to each user who has personal habits when he is working with the Web. This paper focuses on the generalisation of this problem and proposes a multiagent algorithm that, dynamically and according to each user's specific behaviours, computes the personal interest assessment of each of them. Several experimentations show the efficiency of this algorithm, able to self-adapt its functionality in real time when the user habits evolve, compared to existing methods.
Year
DOI
Venue
2010
10.1007/978-3-642-12384-9_2
ADVANCES IN PRACTICAL APPLICATIONS OF AGENTS AND MULTIAGENT SYSTEMS
Keywords
Field
DocType
real time
Generalization,Computer science,A priori and a posteriori,Human–computer interaction,Artificial intelligence,Delegation (computing),Real time learning
Conference
Volume
ISSN
Citations 
70
1867-5662
5
PageRank 
References 
Authors
0.51
8
3
Name
Order
Citations
PageRank
Sylvain Lemouzy1122.78
Valérie Camps29017.42
Pierre Glize321534.04