Clustering the tagged web

Ramage, Daniel; Heymann, Paul; Manning, Christopher D.; García-Molina, Héctor

doi:10.1145/1498759.1498809

Cited by 182 publications

(108 citation statements)

References 26 publications

Supporting

Mentioning

100

Contrasting

Order By: Relevance

“…Similar results can be found elsewhere, including [9], [7], and [37]. Along this line, there have been some recent efforts to identify clusters (topics) of tags and resources, including: [26], which mined tag-based topics via association rules; [42], which iteratively determines user interests and resource topics over a bipartite graph where users and resource are nodes and tagging counts are edges; and [35], which identifies groups of web resources by clustering them via their tags using both K-means and probabilistic clustering. These results motivate our interest in uncovering hidden communities that could help explain these phenomena.…”

Section: Overviewsupporting

confidence: 76%

“…Social bookmarking systems, however, aggregate what would appear to be the independent and uncoordinated tagging actions of a large and heterogeneous tagger population, meaning that it is not obvious that communities of users exist or are detectable. Given the strong evidence of community in the other areas of the Social Web and recent research that has indicated the evidence of coherent tag-based and resourcebased clusters in social bookmarking systems [35,8,41,44,34], we are interested to explore: (i) if user-based communities manifest themselves in social bookmarking systems and how to model them; and (ii) whether this community-based perspective can enhance how users explore the web of socially tagged resources.…”

Section: Figure 1: Two Users Linking To Web Resourcesmentioning

confidence: 99%

See 1 more Smart Citation

Community-based ranking of the social web

Kashoob

Caverlee

Kamath

2010

Proceedings of the 21st ACM Conference on Hypertext and Hypermedia

View full text Add to dashboard Cite

The rise of social interactions on the Web requires developing new methods of information organization and discovery. To that end, we propose a generative community-based probabilistic tagging model that can automatically uncover communities of users and their associated tags. We experimentally validate the quality of the discovered communities over the social bookmarking system Delicious. In comparison to an alternative generative model (Latent Dirichlet Allocation (LDA), we find that the proposed communitybased model improves the empirical likelihood of held-out test data and discovers more coherent interest-based communities. Based on the community-based probabilistic tagging model, we develop a novel community-based ranking model for effective communitybased exploration of socially-tagged Web resources. We compare community-based ranking to three state-of-the-art retrieval models: (i) BM25; (ii) Cluster-based retrieval using K-means clustering; and (iii) LDA-based retrieval. We find that the proposed ranking model results in a significant improvement over these alternatives (from 7% to 22%) in the quality of retrieved pages.

show abstract

Section: Overviewsupporting

confidence: 76%

Section: Figure 1: Two Users Linking To Web Resourcesmentioning

confidence: 99%

Community-based ranking of the social web

Kashoob

Caverlee

Kamath

2010

Proceedings of the 21st ACM Conference on Hypertext and Hypermedia

View full text Add to dashboard Cite

show abstract

“…Greco et al [15] propose a similar approach based on the linear combination of mutual information evaluated on each feature space, where the parameter of the linear combination is automatically determined. Ramage et al [28], propose a generative clustering algorithm based on latent Dirichlet allocation to cluster documents using two different sources of information: document text and tags. Each source is modeled by a probability distribution and a weight value is used to weigh one vector space with respect to the other.…”

Section: Related Workmentioning

confidence: 99%

Parameter-less co-clustering for star-structured heterogeneous data

Ienco

Robardet

Pensa

et al. 2012

Data Min Knowl Disc

View full text Add to dashboard Cite

The availability of data represented with multiple features coming from heterogeneous domains is getting more and more common in real world applications. Such data represent objects of a certain type, connected to other types of data, the features, so that the overall data schema forms a star structure of inter-relationships. Co-clustering these data involves the specification of many parameters, such as the number of clusters for the object dimension and for all the features domains. In this paper we present a novel co-clustering algorithm for heterogeneous star-structured data that is parameter-less. This means that it does not require either the number of row clusters or the number of column clusters for the given feature spaces. Our approach optimizes the Goodman-Kruskal's τ , a measure for cross-association in contingency tables that evaluates the strength of the relationship between two categorical variables. We extend τ to evaluate co-clustering solutions and in particular we apply it in a higher dimensional setting. We propose the algorithm CoStar which optimizes τ by a local search approach. We assess the performance of CoStar on publicly available datasets from the textual and image domains using objective external criteria. The results show that our approach outperforms state-of-the-art methods for the co-clustering of heterogeneous data, while it remains computationally efficient.

show abstract

“…Along this line, [14] incorporates user contributed comments in the data cloud generation process. Some other researchers have explored using user-generated tags and comments for flat clustering, e.g., [18] and [16].…”

Section: Related Workmentioning

confidence: 99%

Hierarchical comments-based clustering

Hsu

Caverlee

Khabiri

2011

Proceedings of the 2011 ACM Symposium on Applied Computing

View full text Add to dashboard Cite

Information resources on the Web like videos, images, and documents are increasingly becoming more "social" through user engagement via commenting systems. These commenting systems provide a forum for users to discuss the resources but have the side effect of providing valuable editorial and contextual information about the resources. In this paper, we explore a comments-driven clustering framework for organizing Web resources according to this user-based perspective. Concretely, we propose a hierarchical comment clustering approach that relies on two key features: (i) comment term normalization and key term extraction for distilling noisy comments for effective clustering; and (ii) a real-time insertion component for incrementally updating the comments-based hierarchy so that resources can be efficiently placed in the hierarchy as comments arise and without the need to re-generate the (potentially) expensive hierarchy. We study the clustering approach over the popular video sharing site YouTube. YouTube is a challenging and difficult environment, notorious for its extremely short, illformed, and often unintelligible user-contributed comments. Through extensive experimental study, we find that the proposed approach can lead to effective and efficient comments-based video organizing even in a YouTube-like environment.

show abstract

Clustering the tagged web

Cited by 182 publications

References 26 publications

Community-based ranking of the social web

Community-based ranking of the social web

Parameter-less co-clustering for star-structured heterogeneous data

Hierarchical comments-based clustering

Contact Info

Product

Resources

About