|
Title:
|
PAGES RECOMMENDATION IN SOCIAL NETWORKS BY CONSIDERING USER AND DOMAIN KNOWLEDGE |
|
Author(s):
|
Hamed Jafarpour, Ahmad A. Kardan |
|
ISBN:
|
978-989-8704-10-8 |
|
Editors:
|
Ajith P. Abraham, Antonio Palma dos Reis and Jörg Roth |
|
Year:
|
2014 |
|
Edition:
|
Single |
|
Keywords:
|
Hybrid recommender system, recommender system, page recommendation, social network, personalized PageRank. |
|
Type:
|
Full Paper |
|
First Page:
|
47 |
|
Last Page:
|
57 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Nowadays, social networks are becoming tremendously popular and many people share their information in the networks with various goals. Generally, the information is published as pages. Whereas many pages are created in social networks, users can not find their favorite pages easily. In this paper, we introduce a new five-dimensional hybrid recommender system which combines collaborative, content-based and demographic filtering techniques and our proposed personalized PageRank. The five dimensions of our proposed system include social relation among users, users profile, liked pages, events and groups which demonstrate users attitudes. Hence, we recommend pages to a user in social networks by considering users attitudes. We consider interests, activities, sex, age, education, work, favourite books, movies, music, sports, TV, athletes, teams, and games which are considered as user knowledge; moreover, social relation among users, iked pages, events and groups are taken into account as domain knowledge to model recommendations. The model attempts to alight influence of cold start problem and synonymy. Finally, the proposed model is applied on three real dataset of Facebook to evaluate and validate the accuracy of recommendations. The results reveal that the accuracy of pages recommendation is obtained as 91% by our proposed model. |
|
|
|
|
|
|
