Dissimilarity measures for histogram-valued observations

Jaejik Kim, L. Billard

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Contemporary datasets can be immense and complex in nature. Thus, summarizing and extracting information frequently precedes any analysis. The summarizing techniques are many and varied and driven by underlying scientific questions of interest. One type of resulting datasets contains so-called histogram-valued observations. While such datasets are becoming more and more pervasive, methodologies to analyse them are still very inadequate. One area of interest falls under the rubric of cluster analysis. Unfortunately, to date, no dis/similarity or distance measures that are readily computable exist for multivariate histogramvalued data. To redress that problem, the present article introduces various dissimilarity measures for histogram data. In particular, extensions to the Gowda-Diday and Ichino-Yaguchi measures for interval data are introduced, along with extensions of some DeCarvalho measures. In addition, a cumulative distribution measure is developed for histograms. These new measures are illustrated for the Fisher iris data and applied to a U.S. temperature dataset.

Original languageEnglish (US)
Pages (from-to)283-303
Number of pages21
JournalCommunications in Statistics - Theory and Methods
Volume42
Issue number2
DOIs
StatePublished - Jan 11 2013

Fingerprint

Dissimilarity Measure
Histogram
Interval Data
Iris
Multivariate Data
Distance Measure
Cluster Analysis
Similarity Measure
Observation
Methodology

Keywords

  • Cumulative distribution dissimilarity measures
  • Extended DeCarvalho
  • Extended Extended Gowda-Diday
  • Ichino-Yaguchi
  • Intersection
  • Iris data
  • Union

ASJC Scopus subject areas

  • Statistics and Probability

Cite this

Dissimilarity measures for histogram-valued observations. / Kim, Jaejik; Billard, L.

In: Communications in Statistics - Theory and Methods, Vol. 42, No. 2, 11.01.2013, p. 283-303.

Research output: Contribution to journalArticle

Kim, Jaejik ; Billard, L. / Dissimilarity measures for histogram-valued observations. In: Communications in Statistics - Theory and Methods. 2013 ; Vol. 42, No. 2. pp. 283-303.
@article{65bbf939c5834928b894efd937ed7b55,
title = "Dissimilarity measures for histogram-valued observations",
abstract = "Contemporary datasets can be immense and complex in nature. Thus, summarizing and extracting information frequently precedes any analysis. The summarizing techniques are many and varied and driven by underlying scientific questions of interest. One type of resulting datasets contains so-called histogram-valued observations. While such datasets are becoming more and more pervasive, methodologies to analyse them are still very inadequate. One area of interest falls under the rubric of cluster analysis. Unfortunately, to date, no dis/similarity or distance measures that are readily computable exist for multivariate histogramvalued data. To redress that problem, the present article introduces various dissimilarity measures for histogram data. In particular, extensions to the Gowda-Diday and Ichino-Yaguchi measures for interval data are introduced, along with extensions of some DeCarvalho measures. In addition, a cumulative distribution measure is developed for histograms. These new measures are illustrated for the Fisher iris data and applied to a U.S. temperature dataset.",
keywords = "Cumulative distribution dissimilarity measures, Extended DeCarvalho, Extended Extended Gowda-Diday, Ichino-Yaguchi, Intersection, Iris data, Union",
author = "Jaejik Kim and L. Billard",
year = "2013",
month = "1",
day = "11",
doi = "10.1080/03610926.2011.581785",
language = "English (US)",
volume = "42",
pages = "283--303",
journal = "Communications in Statistics - Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "2",

}

TY - JOUR

T1 - Dissimilarity measures for histogram-valued observations

AU - Kim, Jaejik

AU - Billard, L.

PY - 2013/1/11

Y1 - 2013/1/11

N2 - Contemporary datasets can be immense and complex in nature. Thus, summarizing and extracting information frequently precedes any analysis. The summarizing techniques are many and varied and driven by underlying scientific questions of interest. One type of resulting datasets contains so-called histogram-valued observations. While such datasets are becoming more and more pervasive, methodologies to analyse them are still very inadequate. One area of interest falls under the rubric of cluster analysis. Unfortunately, to date, no dis/similarity or distance measures that are readily computable exist for multivariate histogramvalued data. To redress that problem, the present article introduces various dissimilarity measures for histogram data. In particular, extensions to the Gowda-Diday and Ichino-Yaguchi measures for interval data are introduced, along with extensions of some DeCarvalho measures. In addition, a cumulative distribution measure is developed for histograms. These new measures are illustrated for the Fisher iris data and applied to a U.S. temperature dataset.

AB - Contemporary datasets can be immense and complex in nature. Thus, summarizing and extracting information frequently precedes any analysis. The summarizing techniques are many and varied and driven by underlying scientific questions of interest. One type of resulting datasets contains so-called histogram-valued observations. While such datasets are becoming more and more pervasive, methodologies to analyse them are still very inadequate. One area of interest falls under the rubric of cluster analysis. Unfortunately, to date, no dis/similarity or distance measures that are readily computable exist for multivariate histogramvalued data. To redress that problem, the present article introduces various dissimilarity measures for histogram data. In particular, extensions to the Gowda-Diday and Ichino-Yaguchi measures for interval data are introduced, along with extensions of some DeCarvalho measures. In addition, a cumulative distribution measure is developed for histograms. These new measures are illustrated for the Fisher iris data and applied to a U.S. temperature dataset.

KW - Cumulative distribution dissimilarity measures

KW - Extended DeCarvalho

KW - Extended Extended Gowda-Diday

KW - Ichino-Yaguchi

KW - Intersection

KW - Iris data

KW - Union

UR - http://www.scopus.com/inward/record.url?scp=84872044029&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84872044029&partnerID=8YFLogxK

U2 - 10.1080/03610926.2011.581785

DO - 10.1080/03610926.2011.581785

M3 - Article

VL - 42

SP - 283

EP - 303

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 2

ER -