Heat Map Visualizations Allow Comparison of Multiple Clustering Results and Evaluation of Dataset Quality: Application to Microarray Data

Sharko, John; Grinstein, Georges; Marx, Kenneth A.; Zhou, Jianping; Cheng, Chia-Ho; Odelberg, Shannon J.; Simon, Hans Georg

doi:10.1109/iv.2007.61

Cited by 14 publications

(9 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The analyst may direct the work of the algorithm, for instance, the training of a self-organizing map [21]. Interactive visualization may allow the analyst to compare results of several runs of clustering and investigate the sensitivity to parameters [23]. Clustering techniques are often included in visualization systems and toolkits, so that the analyst may, on the one hand, use visualization for examining and interpreting results of clustering, on the other hand, use results of clustering for further analysis by means of interactive visual techniques [14] [22].…”

Section: Related Workmentioning

confidence: 99%

Interactive visual clustering of large collections of trajectories

Andrienko

Rinzivillo

Nanni

et al. 2009

2009 IEEE Symposium on Visual Analytics Science and Technology

174

134

View full text Add to dashboard Cite

One of the most common operations in exploration and analysis of various kinds of data is clustering, i.e. discovery and interpretation of groups of objects having similar properties and/or behaviors. In clustering, objects are often treated as points in multi-dimensional space of properties. However, structurally complex objects, such as trajectories of moving entities and other kinds of spatiotemporal data, cannot be adequately represented in this manner. Such data require sophisticated and computationally intensive clustering algorithms, which are very hard to scale effectively to large datasets not fitting in the computer main memory. We propose an approach to extracting meaningful clusters from large databases by combining clustering and classification, which are driven by a human analyst through an interactive visual interface.

show abstract

Section: Related Workmentioning

confidence: 99%

Interactive visual clustering of large collections of trajectories

Andrienko

Rinzivillo

Nanni

et al. 2009

2009 IEEE Symposium on Visual Analytics Science and Technology

174

134

View full text Add to dashboard Cite

show abstract

“…An interactive dissimilarity matrix, presented by Bezdek and Hathaway [Bezdek and Hathaway 2002], was extended to analyze clustering results at different similarity level by Siirtola [Siirtola 2004]. Sharko et al [Sharko et al 2007] use heat maps called cluster stability matrices to visually analyze and reveal most 'stable' clusters in clustering results. In our approach, we enhance the cluster comparison capability in the above studies [Seo and Shneiderman 2002] [Sharko et al 2008] by using IVA operations in the parallel cluster view.…”

Section: Related Workmentioning

confidence: 99%

Integrating cluster formation and cluster evaluation in interactive visual analysis

Turkay

Parulek

Reuter

et al. 2011

Proceedings of the 27th Spring Conference on Computer Graphics

View full text Add to dashboard Cite

Cluster analysis is a popular method for data investigation where data items are structured into groups called clusters. This analysis involves two sequential steps, namely cluster formation and cluster evaluation. In this paper, we propose the tight integration of cluster formation and cluster evaluation in interactive visual analysis in order to overcome the challenges that relate to the black-box nature of clustering algorithms. We present our conceptual framework in the form of an interactive visual environment. In this realization of our framework, we build upon general concepts such as cluster comparison, clustering tendency, cluster stability and cluster coherence. Additionally, we showcase our framework on the cluster analysis of mixed lipid bilayers.

show abstract

“…There are a few approaches to visualize measured similarity values between clusters (or items) in different clustering results instead of explicitly visualizing shared items among multiple clustering results. Sharko et al [ 12 ] utilized a color-coded similarity matrix view to show the stability between items or clusters across different clustering results. Similarities were measured by counting how many times each pair of items was clustered together or how many items each pair of clusters shared.…”

Section: Introductionmentioning

confidence: 99%

“…Color is a powerful visual cue for representing a cluster membership. It is used in many visualization techniques, including parallel coordinate plot [ 6 , 12 ] and scatterplot [ 15 - 17 ], to discriminate clusters while revealing trends in raw data. Similar efforts exist in the visualizations of multiple clustering results.…”

Section: Introductionmentioning

confidence: 99%

XCluSim: a visual analytics tool for interactively comparing multiple clustering results of bioinformatics data

L’Yi

Shin

et al. 2015

BMC Bioinformatics

View full text Add to dashboard Cite

BackgroundThough cluster analysis has become a routine analytic task for bioinformatics research, it is still arduous for researchers to assess the quality of a clustering result. To select the best clustering method and its parameters for a dataset, researchers have to run multiple clustering algorithms and compare them. However, such a comparison task with multiple clustering results is cognitively demanding and laborious.ResultsIn this paper, we present XCluSim, a visual analytics tool that enables users to interactively compare multiple clustering results based on the Visual Information Seeking Mantra. We build a taxonomy for categorizing existing techniques of clustering results visualization in terms of the Gestalt principles of grouping. Using the taxonomy, we choose the most appropriate interactive visualizations for presenting individual clustering results from different types of clustering algorithms. The efficacy of XCluSim is shown through case studies with a bioinformatician.ConclusionsCompared to other relevant tools, XCluSim enables users to compare multiple clustering results in a more scalable manner. Moreover, XCluSim supports diverse clustering algorithms and dedicated visualizations and interactions for different types of clustering results, allowing more effective exploration of details on demand. Through case studies with a bioinformatics researcher, we received positive feedback on the functionalities of XCluSim, including its ability to help identify stably clustered items across multiple clustering results.

show abstract

Heat Map Visualizations Allow Comparison of Multiple Clustering Results and Evaluation of Dataset Quality: Application to Microarray Data

Cited by 14 publications

References 1 publication

Interactive visual clustering of large collections of trajectories

Interactive visual clustering of large collections of trajectories

Integrating cluster formation and cluster evaluation in interactive visual analysis

XCluSim: a visual analytics tool for interactively comparing multiple clustering results of bioinformatics data

Contact Info

Product

Resources

About