Clustme

A visual quality measure for ranking monochrome scatterplots based on cluster patterns

Research output: Contribution to journalArticle

Abstract

We propose ClustMe, a new visual quality measure to rank monochrome scatterplots based on cluster patterns. ClustMe is based on data collected from a human-subjects study, in which 34 participants judged synthetically generated cluster patterns in 1000 scatterplots. We generated these patterns by carefully varying the free parameters of a simple Gaussian Mixture Model with two components. and asked the participants to count the number of clusters they could see (1 or more than 1). Based on the results, we form ClustMe by selecting the model that best predicts these human judgments among 7 different state-of-the-art merging techniques (DEMP). To quantitatively evaluate ClustMe, we conducted a second study, in which 31 human subjects ranked 435 pairs of scatterplots of real and synthetic data in terms of cluster patterns complexity. We use this data to compare ClustMe’s performance to 4 other state-of-the-art clustering measures, including the well-known Clumpiness scagnostics. We found that of all measures, ClustMe is in strongest agreement with the human rankings.

Original languageEnglish
Pages (from-to)225-236
Number of pages12
JournalComputer Graphics Forum
Volume38
Issue number3
DOIs
Publication statusPublished - 1 Jan 2019

Fingerprint

Merging

Keywords

  • Computing methodologies → Cluster analysis
  • Empirical studies in visualization
  • Human-centered computing → Visual analytics
  • Mixture modeling

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design

Cite this

Clustme : A visual quality measure for ranking monochrome scatterplots based on cluster patterns. / Abbas, Mostafa; Aupetit, Michael; Sedlmair, Michael; Bensmail, Halima.

In: Computer Graphics Forum, Vol. 38, No. 3, 01.01.2019, p. 225-236.

Research output: Contribution to journalArticle

@article{d26ca922a887473d84d9d0c7e5454087,
title = "Clustme: A visual quality measure for ranking monochrome scatterplots based on cluster patterns",
abstract = "We propose ClustMe, a new visual quality measure to rank monochrome scatterplots based on cluster patterns. ClustMe is based on data collected from a human-subjects study, in which 34 participants judged synthetically generated cluster patterns in 1000 scatterplots. We generated these patterns by carefully varying the free parameters of a simple Gaussian Mixture Model with two components. and asked the participants to count the number of clusters they could see (1 or more than 1). Based on the results, we form ClustMe by selecting the model that best predicts these human judgments among 7 different state-of-the-art merging techniques (DEMP). To quantitatively evaluate ClustMe, we conducted a second study, in which 31 human subjects ranked 435 pairs of scatterplots of real and synthetic data in terms of cluster patterns complexity. We use this data to compare ClustMe’s performance to 4 other state-of-the-art clustering measures, including the well-known Clumpiness scagnostics. We found that of all measures, ClustMe is in strongest agreement with the human rankings.",
keywords = "Computing methodologies → Cluster analysis, Empirical studies in visualization, Human-centered computing → Visual analytics, Mixture modeling",
author = "Mostafa Abbas and Michael Aupetit and Michael Sedlmair and Halima Bensmail",
year = "2019",
month = "1",
day = "1",
doi = "10.1111/cgf.13684",
language = "English",
volume = "38",
pages = "225--236",
journal = "Computer Graphics Forum",
issn = "0167-7055",
publisher = "Wiley-Blackwell",
number = "3",

}

TY - JOUR

T1 - Clustme

T2 - A visual quality measure for ranking monochrome scatterplots based on cluster patterns

AU - Abbas, Mostafa

AU - Aupetit, Michael

AU - Sedlmair, Michael

AU - Bensmail, Halima

PY - 2019/1/1

Y1 - 2019/1/1

N2 - We propose ClustMe, a new visual quality measure to rank monochrome scatterplots based on cluster patterns. ClustMe is based on data collected from a human-subjects study, in which 34 participants judged synthetically generated cluster patterns in 1000 scatterplots. We generated these patterns by carefully varying the free parameters of a simple Gaussian Mixture Model with two components. and asked the participants to count the number of clusters they could see (1 or more than 1). Based on the results, we form ClustMe by selecting the model that best predicts these human judgments among 7 different state-of-the-art merging techniques (DEMP). To quantitatively evaluate ClustMe, we conducted a second study, in which 31 human subjects ranked 435 pairs of scatterplots of real and synthetic data in terms of cluster patterns complexity. We use this data to compare ClustMe’s performance to 4 other state-of-the-art clustering measures, including the well-known Clumpiness scagnostics. We found that of all measures, ClustMe is in strongest agreement with the human rankings.

AB - We propose ClustMe, a new visual quality measure to rank monochrome scatterplots based on cluster patterns. ClustMe is based on data collected from a human-subjects study, in which 34 participants judged synthetically generated cluster patterns in 1000 scatterplots. We generated these patterns by carefully varying the free parameters of a simple Gaussian Mixture Model with two components. and asked the participants to count the number of clusters they could see (1 or more than 1). Based on the results, we form ClustMe by selecting the model that best predicts these human judgments among 7 different state-of-the-art merging techniques (DEMP). To quantitatively evaluate ClustMe, we conducted a second study, in which 31 human subjects ranked 435 pairs of scatterplots of real and synthetic data in terms of cluster patterns complexity. We use this data to compare ClustMe’s performance to 4 other state-of-the-art clustering measures, including the well-known Clumpiness scagnostics. We found that of all measures, ClustMe is in strongest agreement with the human rankings.

KW - Computing methodologies → Cluster analysis

KW - Empirical studies in visualization

KW - Human-centered computing → Visual analytics

KW - Mixture modeling

UR - http://www.scopus.com/inward/record.url?scp=85070056315&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070056315&partnerID=8YFLogxK

U2 - 10.1111/cgf.13684

DO - 10.1111/cgf.13684

M3 - Article

VL - 38

SP - 225

EP - 236

JO - Computer Graphics Forum

JF - Computer Graphics Forum

SN - 0167-7055

IS - 3

ER -