Finding Cohesive Clusters for Analyzing Knowledge Communities

Kandylas, Vasileios; Upham, S. Phineas; Ungar, Lyle H.

doi:10.1109/icdm.2007.22

Cited by 7 publications

(4 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Drawing all time stamps from the same beta distribution might be not appropriate for, such as, stream data [39]. Some other approaches are, for instance, He et al [16] develop inheritance topic model to understand topic evolution by leveraging the citation information; Kandylas et al [19] analyse the evolution of knowledge communities based on the clustering over time method, called Streemer.…”

Section: Related Workmentioning

confidence: 99%

Sequential latent Dirichlet allocation

Buntine

Jin

et al. 2011

Knowl Inf Syst

View full text Add to dashboard Cite

Understanding how topics within a document evolve over the structure of the document is an interesting and potentially important problem in exploratory and predictive text analytics. In this article, we address this problem by presenting a novel variant of latent Dirichlet allocation (LDA): Sequential LDA (SeqLDA). This variant directly considers the underlying sequential structure, i.e. a document consists of multiple segments (e.g. chapters, paragraphs), each of which is correlated to its antecedent and subsequent segments. Such progressive sequential dependency is captured by using the hierarchical two-parameter Poisson-Dirichlet process (HPDP). We develop an efficient collapsed Gibbs sampling algorithm to sample from the posterior of the SeqLDA based on the HPDP. Our experimental results on patent documents show that by considering the sequential structure within a document, our SeqLDA model has a higher fidelity over LDA in terms of perplexity (a standard measure of dictionary-based compressibility). The SeqLDA model also yields a nicer sequential topic structure than LDA, as we show in experiments on several books such as Melville's 'Moby Dick'.

show abstract

Section: Related Workmentioning

confidence: 99%

Sequential latent Dirichlet allocation

Buntine

Jin

et al. 2011

Knowl Inf Syst

View full text Add to dashboard Cite

show abstract

“…Morris (2005) proposed a model to monitor the birth and development of a scientific speciality using a collection of journal papers. Kandylas, Upham, and Ungar (2008) intensively analyzed how communities of researchers are evolved over time and designed the model for community growth. Zhou, Councill, Zha, and Giles (2007) also proposed a method to discover the temporal trends of researchers by constructing their co-authorship graphs over time and comparing their communities.…”

Section: Related Workmentioning

confidence: 99%

Network analysis of temporal trends in scholarly research productivity

Kim

Yoon

Crowcroft

2012

Journal of Informetrics

View full text Add to dashboard Cite

“…This metric can be used to evaluate clusters and the partition in many applications. For example, while Zhong and Ghosh (2005) used NMI for evaluating clusters in document clustering, (Kandylas et al, 2008) used it for community knowledge analysis and (Long et al, 2010) used it for evaluating graph clustering. Hadjitodorov et al (2006) proposed a selective strategy which is based on diversity.…”

Section: Literature Review 21 Cluster Ensemblementioning

confidence: 99%

Wisdom of Crowds cluster ensemble

Alizadeh

Yousefnezhad

Minaei-Bidgoli

2015

IDA

View full text Add to dashboard Cite

Abstract:The Wisdom of Crowds is a phenomenon described in social science that suggests four criteria applicable to groups of people. It is claimed that, if these criteria are satisfied, then the aggregate decisions made by a group will often be better than those of its individual members.Inspired by this concept, we present a novel feedback framework for the cluster ensemble problem, which we call Wisdom of Crowds Cluster Ensemble (WOCCE). Although many conventional cluster ensemble methods focusing on diversity have recently been proposed, WOCCE analyzes the conditions necessary for a crowd to exhibit this collective wisdom. These include decentralization criteria for generating primary results, independence criteria for the base algorithms, and diversity criteria for the ensemble members. We suggest appropriate procedures for evaluating these measures, and propose a new measure to assess the diversity. We evaluate the performance of WOCCE against some other traditional base algorithms as well as state-of-the-art ensemble methods. The results demonstrate the efficiency of WOCCE's aggregate decisionmaking compared to other algorithms.

show abstract

Finding Cohesive Clusters for Analyzing Knowledge Communities

Cited by 7 publications

References 17 publications

Sequential latent Dirichlet allocation

Sequential latent Dirichlet allocation

Network analysis of temporal trends in scholarly research productivity

Wisdom of Crowds cluster ensemble

Contact Info

Product

Resources

About