Multidimensional clustering based collaborative filtering approach for diversified recommendation

Xiaohui Li, Tomohiro Murata

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

This paper presents a hybrid recommendation approach that is used for discovering potential information with multidimensional clustering in recommender systems. This facilitates to obtain user groups for improving effectiveness and diversity of recommendation. The proposed algorithm works in three phases. In first phase, user groups are collected in the form of user profile, which applied multidimensional clustering algorithm and stored in the database for future recommendation. In second phase, the appropriate clusters are chosen using pruning of clusters. In third phase, the recommendations are generated for target user with similarity measures and quality rating prediction. The performance of proposed approach is evaluated using a public movie dataset and compared with two representative recommendation algorithms. The experimental results demonstrate that our proposed approach performs superiorly and alleviates problems, such as cold-start and data sparsity in collaborative filtering recommendation.

Original languageEnglish
Title of host publicationICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education
Pages905-910
Number of pages6
DOIs
Publication statusPublished - 2012
Event2012 7th International Conference on Computer Science and Education, ICCSE 2012 - Melbourne, VIC
Duration: 2012 Jul 142012 Jul 17

Other

Other2012 7th International Conference on Computer Science and Education, ICCSE 2012
CityMelbourne, VIC
Period12/7/1412/7/17

Fingerprint

Collaborative filtering
Collaborative Filtering
Recommendations
Clustering
Recommender systems
Clustering algorithms
User Profile
Recommender Systems
Pruning
Sparsity
Similarity Measure
Clustering Algorithm
Target
Prediction
Experimental Results
Demonstrate

Keywords

  • clustering
  • collaborative filtering
  • multidimensional data
  • recommender systems

ASJC Scopus subject areas

  • Computer Science (miscellaneous)
  • Theoretical Computer Science

Cite this

Li, X., & Murata, T. (2012). Multidimensional clustering based collaborative filtering approach for diversified recommendation. In ICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education (pp. 905-910). [6295214] https://doi.org/10.1109/ICCSE.2012.6295214

Multidimensional clustering based collaborative filtering approach for diversified recommendation. / Li, Xiaohui; Murata, Tomohiro.

ICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education. 2012. p. 905-910 6295214.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, X & Murata, T 2012, Multidimensional clustering based collaborative filtering approach for diversified recommendation. in ICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education., 6295214, pp. 905-910, 2012 7th International Conference on Computer Science and Education, ICCSE 2012, Melbourne, VIC, 12/7/14. https://doi.org/10.1109/ICCSE.2012.6295214
Li X, Murata T. Multidimensional clustering based collaborative filtering approach for diversified recommendation. In ICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education. 2012. p. 905-910. 6295214 https://doi.org/10.1109/ICCSE.2012.6295214
Li, Xiaohui ; Murata, Tomohiro. / Multidimensional clustering based collaborative filtering approach for diversified recommendation. ICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education. 2012. pp. 905-910
@inproceedings{74987ca3cb7a4235a8cfbfe9ca24a2c7,
title = "Multidimensional clustering based collaborative filtering approach for diversified recommendation",
abstract = "This paper presents a hybrid recommendation approach that is used for discovering potential information with multidimensional clustering in recommender systems. This facilitates to obtain user groups for improving effectiveness and diversity of recommendation. The proposed algorithm works in three phases. In first phase, user groups are collected in the form of user profile, which applied multidimensional clustering algorithm and stored in the database for future recommendation. In second phase, the appropriate clusters are chosen using pruning of clusters. In third phase, the recommendations are generated for target user with similarity measures and quality rating prediction. The performance of proposed approach is evaluated using a public movie dataset and compared with two representative recommendation algorithms. The experimental results demonstrate that our proposed approach performs superiorly and alleviates problems, such as cold-start and data sparsity in collaborative filtering recommendation.",
keywords = "clustering, collaborative filtering, multidimensional data, recommender systems",
author = "Xiaohui Li and Tomohiro Murata",
year = "2012",
doi = "10.1109/ICCSE.2012.6295214",
language = "English",
isbn = "9781467302425",
pages = "905--910",
booktitle = "ICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education",

}

TY - GEN

T1 - Multidimensional clustering based collaborative filtering approach for diversified recommendation

AU - Li, Xiaohui

AU - Murata, Tomohiro

PY - 2012

Y1 - 2012

N2 - This paper presents a hybrid recommendation approach that is used for discovering potential information with multidimensional clustering in recommender systems. This facilitates to obtain user groups for improving effectiveness and diversity of recommendation. The proposed algorithm works in three phases. In first phase, user groups are collected in the form of user profile, which applied multidimensional clustering algorithm and stored in the database for future recommendation. In second phase, the appropriate clusters are chosen using pruning of clusters. In third phase, the recommendations are generated for target user with similarity measures and quality rating prediction. The performance of proposed approach is evaluated using a public movie dataset and compared with two representative recommendation algorithms. The experimental results demonstrate that our proposed approach performs superiorly and alleviates problems, such as cold-start and data sparsity in collaborative filtering recommendation.

AB - This paper presents a hybrid recommendation approach that is used for discovering potential information with multidimensional clustering in recommender systems. This facilitates to obtain user groups for improving effectiveness and diversity of recommendation. The proposed algorithm works in three phases. In first phase, user groups are collected in the form of user profile, which applied multidimensional clustering algorithm and stored in the database for future recommendation. In second phase, the appropriate clusters are chosen using pruning of clusters. In third phase, the recommendations are generated for target user with similarity measures and quality rating prediction. The performance of proposed approach is evaluated using a public movie dataset and compared with two representative recommendation algorithms. The experimental results demonstrate that our proposed approach performs superiorly and alleviates problems, such as cold-start and data sparsity in collaborative filtering recommendation.

KW - clustering

KW - collaborative filtering

KW - multidimensional data

KW - recommender systems

UR - http://www.scopus.com/inward/record.url?scp=84868098363&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84868098363&partnerID=8YFLogxK

U2 - 10.1109/ICCSE.2012.6295214

DO - 10.1109/ICCSE.2012.6295214

M3 - Conference contribution

SN - 9781467302425

SP - 905

EP - 910

BT - ICCSE 2012 - Proceedings of 2012 7th International Conference on Computer Science and Education

ER -