|
Title:
|
UTILIZING ITEM TYPE INFORMATION TO REMOVE MEANINGLESS DATA IN RECOMMENDATION SYSTEMS |
|
Author(s):
|
Yadong Liu |
|
ISBN:
|
978-972-8939-09-0 |
|
Editors:
|
Miguel Baptista Nunes, Pedro Isaías and Philip Powell |
|
Year:
|
2010 |
|
Edition:
|
Single |
|
Keywords:
|
Meaningless data, algorithm, major types based |
|
Type:
|
Short Paper |
|
First Page:
|
367 |
|
Last Page:
|
371 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Personalized recommendation systems are web-based systems that aim at predicting a user's interests on available products and services by relying on previously rated items and information. One of the most commonly used and successful recommendation approaches is collaborative filtering, which explores the correlations within user-item interactions to infer user interests and preferences. However, the prediction quality of collaborative filtering is greatly limited by problems such as information explosion. Error elimination turns to be the key point of improving the prediction accuracy because irrelevant and meaningless data contributes to increasing deviation of prediction errors. In this paper, we propose a user-based collaborative filtering approach which applies the information of the rated items to remove irrelevant and meaningless data and thus improve prediction accuracy. The experiments suggest that the new item information based approach contributes to substantial improvement of prediction accuracy, without a meaningful increase in running time. |
|
|
|
|
|
|
