The week's pick
Reseach Article
YouTube Video Recommendation via Cross-Network Collaboration
| International Journal of Computer Applications |
| Foundation of Computer Science (FCS), NY, USA |
| Volume 146 - Number 11 |
| Year of Publication: 2016 |
| Authors: Tejal T. Wanave, Sheetal A. Takale |
10.5120/ijca2016910896
|
Tejal T. Wanave, Sheetal A. Takale . YouTube Video Recommendation via Cross-Network Collaboration. International Journal of Computer Applications. 146, 11 ( Jul 2016), 9-17. DOI=10.5120/ijca2016910896
Abstract
We present a system that automatically recommends videos to the three YouTube users: new, light and heavy. We have proposed YouTube video cross network recommendation system; it extracts users auxiliary information on Twitter to address the three typical problems: new user, cold start and sparsity which are occurring in single network recommendation system. At the first stage, the system recommends videos to the new user using cross relevance method, it maps individual user Twitter preferences with video titles on YouTube. Our system has calculated a tweet vocabulary of more than 400 words using Hash Map function. At the second stage, we construct textual and visual similarity between user's data on different OSNs (Online Social Networks), i.e. the system recommends videos to light and heavy users by considering their data on both Twitter and YouTube. Here we map visual based features of user- uploaded videos on YouTube with all other videos in a database also uploaded video titles is used to automatically suggest relevant videos on the basis of text similarity. These methods give benefit to new, light and heavy users who are having limited or fr++++ equent behavior record on YouTube. Finally, we compared our cross-relevance method with other single network based methods a) the average relevance of videos automatically recommended by our system for new YouTube users is 76% with Top K=5. b) The average relevance of videos automatically recommended by our system for light and heavy YouTube users is 90% with Top K=5.
References
- Zhengyu Deng, Jita Sang, and Changsheng Xu, “Unified YouTube Video Recommendation via Cross network Collaboration," June 23–26, 2015.
- http://nlprweb.ia.ac.cn/mmc/homepage/myan/Project_YanMing/ming_ICMR2015/crossnetwork%20video%20recommendation.html.
- Mi Zhang, Jie Tang, “Addressing Cold Start in Recommender Systems: A Semi-supervised Co-training Algorithm,”In Shanghai Key Laboratory of Intelligent Information Processing, Copyright 2014 ACM 978-1-4503-2257-7/14/07, July 6–11, 2014.
- Reza Zafarani and Huan Liu, “Connecting users across social media sites: a behavioral-modeling approach,” In ACM SIGKDD 2013, pages 41–49, ACM.
- Zhengyu Deng, Jitao Sang, “Personalized Video recommendation based on cross-platform user modeling,” In ICME 2013, pages 1–6, IEEE.
- Suman Deb Roy, Tao Mei, Wenjun Zeng, “Social transfer: cross-domain transfer learning from social streams for media application,” In ACM Multimedia 2012, pages 649–658.
- ChanLe Wu, Ming Xie, “Solving the Sparsity Problem in Recommender Systems Using Association Retrieval, “In National Engineering Research Center for Multimedia Software, Wuhan, China, © 2011 ACADEMY PUBLISHER.
- James Davidson, Benjamin Liebald, “The YouTube video recommendation System,” In Proceedings of the fourth ACM conference on Recommender systems, pages 293–296, 2010.
- Hao Ma, “Recommender systems with social regularization,” In Proceedings of the fourth ACM international conference on Web search and data mining, pages 287–296, ACM, 2011.
- Fabian Abel, “Analyzing cross-system user modeling on the social web,” In Web Engineering, pages 28–43, Springer, 2011.
- Zi-Ke Zhang, Chuang Liu, “Solving the cold-start problem in recommender systems with social tags,” EPL (Euro physics Letters), 92(2):28002, 2010.
- James Davidson, Benjamin Liebald, “The YouTube video recommendation system,” In Proceedings of the fourth ACM conference on Recommender systems, pages 293–296, ACM, 2010.
- Yehuda Koren, “Factorization meets the neighborhood: a Multifaceted collaborative filtering model,” In ACM 2008. Pages426–434.
Index Terms
Keywords