Comparing Fuzzy Partitions: A Generalization of the Rand Index and Related Measures

Hüllermeier, Eyke; Rifqi, Maria; Henzgen, Sascha; Senge, Robin

doi:10.1109/tfuzz.2011.2179303

Cited by 104 publications

(83 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this setting hard clusters are regarded as a special case of possible soft clusters and will likely be punished by the soft-clustering evaluation method. We use fuzzy BCubed Precision, Recall and F-Measure metrics reported in [9] and [7]. According to the analysis of formal constraints that a cluster evaluation metric needs to fulfill [2], fuzzy BCubed metrics are superior to Purity, Inverse Purity, Mutual Information, Rand Index, etc.…”

Section: Methodsmentioning

confidence: 99%

“…Prior to the summarization step, comments are grouped into clusters using K-Means clustering and LDA. Although the nature of LDA allows soft clustering 6 , the authors convert LDA output to hard-clusters 7 by assigning a comment C to the most likely topic, i.e. the topic t r that maximizes P (C|t r ) * P (t r ), where r is the topic/cluster index.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Graph-Based Approach to Topic Clustering for Online Comments to News

Aker

Kurtić

Balamurali

et al. 2016

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper investigates graph-based approaches to labeled topic clustering of reader comments in online news. For graph-based clustering we propose a linear regression model of similarity between the graph nodes (comments) based on similarity features and weights trained using automatically derived training data. To label the clusters our graph-based approach makes use of DBPedia to abstract topics extracted from the clusters. We evaluate the clustering approach against gold standard data created by human annotators and compare its results against LDA -currently reported as the best method for the news comment clustering task. Evaluation of cluster labelling is set up as a retrieval task, where human annotators are asked to identify the best cluster given a cluster label. Our clustering approach significantly outperforms the LDA baseline and our evaluation of abstract cluster labels shows that graph-based approaches are a promising method of creating labeled clusters of news comments, although we still find cases where the automatically generated abstractive labels are insufficient to allow humans to correctly associate a label with its cluster.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A Graph-Based Approach to Topic Clustering for Online Comments to News

Aker

Kurtić

Balamurali

et al. 2016

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…In the fuzzy framework, one of the most recent and frequently used measure is the Fuzzy Rand Index (FRI) suggested by Hüllermeier et al (2012) and defined as follows:…”

Section: Clustering Algorithmmentioning

confidence: 99%

Copula-based fuzzy clustering of spatial time series

2017

View full text Add to dashboard Cite

This paper contributes to the existing literature on the analysis of spatial time series presenting a new clustering algorithm called COFUST, i.e. COpula-based FUzzy clustering algorithm for Spatial Time series. The underlying idea of this algorithm is to perform a fuzzy Partitioning Around Medoids (PAM) clustering using copula-based approach to interpret comovements of time series. This generalisation allows both to extend usual clustering methods for time series based on Pearson's correlation and to capture the uncertainty that arises assigning units to clusters. Furthermore, its flexibility permits to include directly in the algorithm the spatial information. Our approach is presented and discussed using both simulated and real data, highlighting its main advantages.

show abstract

“…This is because entries of a consensus matrix may appear decimal points. We note that there are several fuzzy extensions of RI in the literature (see [1,7,8,19]). However, our recently proposed fuzzy generalized Rand index [42] is well used as an evaluation measure for cluster ensembles.…”

Section: The Proposed Evaluation Measure For Cluster Ensemblesmentioning

confidence: 99%