Liang Gou
ABSTRACT
Multimedia ranking algorithms are usually user-neutral and measure the importance and relevance of documents by only using the visual contents and meta-data. However, users' interests and preferences are often diverse, and may demand different results even with the same queries. How can we integrate user interests in ranking algorithms to improve search results? Here, we introduce Social Network Document Rank (SNDocRank), a new ranking framework that considers a searcher's social network, and apply it to video search. SNDocRank integrates traditional tf-idf ranking with our Multi-level Actor Similarity (MAS) algorithm, which measures the similarity between social networks of a searcher and document owners. Results from our evaluation study with a social network and video data from YouTube show that SNDocRank offers search results more relevant to user's interests than other traditional ranking methods.
- Barabasi, A. L., & Bonabeau, E. Scale-free Free Networks. Scientific American, 288(5), pp. 50--59, 2003.Google Scholar
Cross Ref
- Bellucci, A., Ghiron, S. L., Aedo, I., & Malizia, A. Visual tag authoring: picture extraction via localized, collaborative tagging. In ACM AVI, pp. 351--354, Napoli, Italy, 2008. Google ScholarDigital Library
- Chang, S.-F. et al, Columbia University TRECVID-2006 video search and high-level feature extraction, NIST TRECVID workshop, 2006.Google Scholar
- Clauset, A., Newman, M. E. J., & Moore, C. Finding community structure in very large networks. Phy. Rev. E, 70(6), 66111, 2004.Google ScholarCross Ref
- Clough, P., Grubinger, M., Deselaers, T., Hanbury, A., & Muller, H. Overview of the imageclef 2007 photographic retrieval task. In CLEF 2007, vol. 5152 of LNCS. CLEF, Springer-Verlag, 2008. Google ScholarDigital Library
- Coppersmith, D. & Winograd, S. Matrix multiplication via arithmetic progressions. Journal of Symbolic Computation, 9:251--280, 1990. Google ScholarDigital Library
- Escalante, H. J., Hérnadez, C. A., Sucar, L. E., & Montes, M. Late fusion of heterogeneous methods for multimedia image retrieval. In ACM MIR, pp. 172--179, Canada, 2008. Google ScholarDigital Library
- Goodrum, A. Image information retrieval: An overview of current research. Journal of Informing Science, 3(2), 2000.Google Scholar
- Gou, L., J. Kim, H. Chen, J. Collins, M. Goodman, X. L. Zhang & C. L. Giles. MobiSNA: a Mobile Video Social Network Application. Proc. of MobiDE'09: 53--56. Google ScholarDigital Library
- Hsu, W. H., Kennedy, L. S., & Chang, S. Video search reranking through random walk over document-level context graph. In ACM Multimedia, pp. 172--179, Germany, 2007. Google ScholarDigital Library
- Jarvelin, K. & Kekalainen, J. IR evaluation methods for retrieving highly relevant documents, Proc. of the 23rd SIGIR conference, pp. 41--48, 2000. Google ScholarDigital Library
- Kennedy, L. S. & Chang, S. A reranking approach for context-based concept fusion in video indexing and retrieval. In ACM CIVR, The Netherlands, pp. 333--340, 2007. Google ScholarDigital Library
- Kennedy, L., Naaman, M., Ahern, S., Nair, R., & Rattenbury, T. How flickr helps us make sense of the world: context and content in community-contributed media collections. ACM Multimedia, pp. 631--640, Germany, 2007. Google ScholarDigital Library
- Landis, R. J. & Koch, G. The measurement of observer agreement for categorical data. Biometrics, 33, pp. 159--174, 1979.Google ScholarCross Ref
- Leicht, E. A., Holme, P., & Newman, M. E. J. Vertex similarity in networks. APS, Vol. 73, pp. 26120, 2006.Google ScholarCross Ref
- Lorrain, F. & White, H. Structural equivalence of individuals in social networks. Math. Sociol, 1, pp. 49--80, 1971.Google ScholarCross Ref
- McPherson, M., Smith-Lovin, L., et al. Birds of a Feather: Homophily in Social Networks. Annual Review of Sociology, 27(415): 444, 2001.Google Scholar
- Moxley, E., Kleban, J., & Manjunath, B. S. Spirittagger: a geo-aware tag suggestion tool mined from flickr. In ACM MIR, pp. 24--30, Canada, 2008. Google ScholarDigital Library
- Newman, M. E. J. & Girvan, M. Finding and evaluating community structure in networks. Phys. Rev. E, 69, pp.26113, 2004.Google ScholarCross Ref
- Rautiainen, M. M. & Seppdnen, T.. Comparison of visual features and fusion techniques in automatic detection of concepts from news video. In Proc. of the IEEE ICME, pp. 932--935, 2005.Google ScholarCross Ref
- Rautiainen, M., Ojala, T., & Tapio, S. Analysing the performance of visual, concept and text features in content-based video retrieval. In ACM MIR, pp.197--204, New York, NY, USA, 2004. Google ScholarDigital Library
- Salton, G. & Buckley, C. Term-weighting approaches in automatic text retrieval. Info. Proc. & Mgt., 24 (5), pp. 513--523, 1988. Google ScholarDigital Library
- Salton, G. Automatic Text Processing: The Transformation, Analysis, and Retrieval of Information by Computer. Addison-Wesley, Reading, MA, 1989. Google ScholarDigital Library
- Smeulders, A. W., Worring, M., Santini, S., Gupta, A., & Jain, R. Content-based image retrieval at the end of the early years. IEEE Trans. Pattern Anal. Mach. Intell. 22(12), pp. 1349--1380, 2000. Google ScholarDigital Library
- Snoek, C., Worring, M., and Smeulders, A. Early versus late fusion in semantic video analysis. In ACM Multimedia, pp. 399--402, Singapore, 2005. Google ScholarDigital Library
- Wasserman, S., & Faust, K. Social network analysis: methods and applications. Cambridge University Press, MA 1994.Google Scholar
- Yan, R., Hauptmann, A., & Jin, R.. Multimedia search with pseudo-relevance feedback. In AAAI Spring Symposium on IMKM, Palo Alto, CA, 2003.Google ScholarCross Ref
- Yang, Y. H., Wu, P. T., Lee, C. W., Lin, K. H., Hsu, W. H., & Chen, H. H. ContextSeer: context search and recommendation at query time for shared consumer photos. In ACM Multimedia, pp. 199--208, Canada, 2008. Google ScholarDigital Library
- Yuan, J., Luo, J., Kautz, H., & Wu, Y. Mining GPS traces and visual words for event classification. In ACM MIR, pp. 2--9, Canada, 2008. Google ScholarDigital Library
Index Terms
- SNDocRank: a social network-based video search ranking framework
Recommendations
Social network document ranking
JCDL '10: Proceedings of the 10th annual joint conference on Digital librariesIn search engines, ranking algorithms measure the importance and relevance of documents mainly based on the contents and relationships between documents. User attributes are usually not considered in ranking. This user-neutral approach, however, may not ...
SNDocRank: document ranking based on social networks
WWW '10: Proceedings of the 19th international conference on World wide webTo improve the search results for socially-connect users, we propose a ranking framework, Social Network Document Rank (SNDocRank). This framework considers both document contents and the similarity between a searcher and document owners in a social ...
Intent and its discontents: the user at the wheel of the online video search engine
MM '12: Proceedings of the 20th ACM international conference on MultimediaWe embrace the position of the user in the driver's seat of the video search engine by proposing a principled framework for multimedia retrieval that moves beyond what users are searching for also to encompass why they search. This 'why' is understood ...
Reviews
Sithu D. Sudarsan
An ongoing challenge for researchers is the ability to provide relevant search results to users. This paper attempts to arrive at the ranking score of search results by using the user's behavior pattern, which is identified from the user's social networking profile. Gou et al. present their case for considering user-specific parameters to arrive at the relevance score, in a simple way. Their framework arrives at the relevancy score based on two aspects: one is dependent on the document corpus and its characteristics, which mainly depend on the values of the popular term frequency (tf) and inverse document frequency (idf); the other aspect is dependent on the user's behavior and his or her contacts on social networking sites such as Facebook, Flickr, and YouTube. The paper presents a case study based on a simulated social network and dataset from YouTube. The authors use SNDocRank, tf-idf, and cosine similarity to evaluate the search result's ranking. They claim that the proposed method ranks results better for a given user. Although the work and results are interesting, the authors fail to present two important additional issues. The first is the ability to analyze the social networking aspect of each user-this may not be always correct, as a user might have multiple social network identifications and interests-and the challenge of gathering specific information to enable social network analysis for each new user. The second issue is the computational/communication overhead that will also add latency in presenting the search results. A quick outline of these issues would have added a lot of value to the paper. Online Computing Reviews Service
Access critical reviews of Computing literature here
Become a reviewer for Computing Reviews.
Comments