High dimensional data analysis using multivariate generalized spatial quantiles

Nitai D. Mukhopadhyay, Singdhansu B Chatterjee

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

High dimensional data routinely arises in image analysis, genetic experiments, network analysis, and various other research areas. Many such datasets do not correspond to well-studied probability distributions, and in several applications the data-cloud prominently displays non-symmetric and non-convex shape features. We propose using spatial quantiles and their generalizations, in particular, the projection quantile, for describing, analyzing and conducting inference with multivariate data. Minimal assumptions are made about the nature and shape characteristics of the underlying probability distribution, and we do not require the sample size to be as high as the data-dimension. We present theoretical properties of the generalized spatial quantiles, and an algorithm to compute them quickly. Our quantiles may be used to obtain multidimensional confidence or credible regions that are not required to conform to a pre-determined shape. We also propose a new notion of multidimensional order statistics, which may be used to obtain multidimensional outliers. Many of the features revealed using a generalized spatial quantile-based analysis would be missed if the data was shoehorned into a well-known probabilistic configuration.

Original languageEnglish (US)
Pages (from-to)768-780
Number of pages13
JournalJournal of Multivariate Analysis
Volume102
Issue number4
DOIs
StatePublished - Apr 1 2011

Fingerprint

High-dimensional Data
Quantile
Probability distributions
Data analysis
Electric network analysis
Image analysis
Statistics
Probability Distribution
Shape Feature
Multivariate Data
Network Analysis
Order Statistics
Image Analysis
Experiments
Outlier
Confidence
Sample Size
Projection
Configuration
Experiment

Keywords

  • Brain imaging
  • Generalized spatial quantile
  • High dimensional data visualization
  • Multidimensional coverage sets
  • Multivariate order statistics
  • Multivariate quantile
  • Projection quantile
  • Spatial quantile

Cite this

High dimensional data analysis using multivariate generalized spatial quantiles. / Mukhopadhyay, Nitai D.; Chatterjee, Singdhansu B.

In: Journal of Multivariate Analysis, Vol. 102, No. 4, 01.04.2011, p. 768-780.

Research output: Contribution to journalArticle

@article{575a32f62b1b4cff8efdb29ace6780fb,
title = "High dimensional data analysis using multivariate generalized spatial quantiles",
abstract = "High dimensional data routinely arises in image analysis, genetic experiments, network analysis, and various other research areas. Many such datasets do not correspond to well-studied probability distributions, and in several applications the data-cloud prominently displays non-symmetric and non-convex shape features. We propose using spatial quantiles and their generalizations, in particular, the projection quantile, for describing, analyzing and conducting inference with multivariate data. Minimal assumptions are made about the nature and shape characteristics of the underlying probability distribution, and we do not require the sample size to be as high as the data-dimension. We present theoretical properties of the generalized spatial quantiles, and an algorithm to compute them quickly. Our quantiles may be used to obtain multidimensional confidence or credible regions that are not required to conform to a pre-determined shape. We also propose a new notion of multidimensional order statistics, which may be used to obtain multidimensional outliers. Many of the features revealed using a generalized spatial quantile-based analysis would be missed if the data was shoehorned into a well-known probabilistic configuration.",
keywords = "Brain imaging, Generalized spatial quantile, High dimensional data visualization, Multidimensional coverage sets, Multivariate order statistics, Multivariate quantile, Projection quantile, Spatial quantile",
author = "Mukhopadhyay, {Nitai D.} and Chatterjee, {Singdhansu B}",
year = "2011",
month = "4",
day = "1",
doi = "10.1016/j.jmva.2010.12.002",
language = "English (US)",
volume = "102",
pages = "768--780",
journal = "Journal of Multivariate Analysis",
issn = "0047-259X",
publisher = "Academic Press Inc.",
number = "4",

}

TY - JOUR

T1 - High dimensional data analysis using multivariate generalized spatial quantiles

AU - Mukhopadhyay, Nitai D.

AU - Chatterjee, Singdhansu B

PY - 2011/4/1

Y1 - 2011/4/1

N2 - High dimensional data routinely arises in image analysis, genetic experiments, network analysis, and various other research areas. Many such datasets do not correspond to well-studied probability distributions, and in several applications the data-cloud prominently displays non-symmetric and non-convex shape features. We propose using spatial quantiles and their generalizations, in particular, the projection quantile, for describing, analyzing and conducting inference with multivariate data. Minimal assumptions are made about the nature and shape characteristics of the underlying probability distribution, and we do not require the sample size to be as high as the data-dimension. We present theoretical properties of the generalized spatial quantiles, and an algorithm to compute them quickly. Our quantiles may be used to obtain multidimensional confidence or credible regions that are not required to conform to a pre-determined shape. We also propose a new notion of multidimensional order statistics, which may be used to obtain multidimensional outliers. Many of the features revealed using a generalized spatial quantile-based analysis would be missed if the data was shoehorned into a well-known probabilistic configuration.

AB - High dimensional data routinely arises in image analysis, genetic experiments, network analysis, and various other research areas. Many such datasets do not correspond to well-studied probability distributions, and in several applications the data-cloud prominently displays non-symmetric and non-convex shape features. We propose using spatial quantiles and their generalizations, in particular, the projection quantile, for describing, analyzing and conducting inference with multivariate data. Minimal assumptions are made about the nature and shape characteristics of the underlying probability distribution, and we do not require the sample size to be as high as the data-dimension. We present theoretical properties of the generalized spatial quantiles, and an algorithm to compute them quickly. Our quantiles may be used to obtain multidimensional confidence or credible regions that are not required to conform to a pre-determined shape. We also propose a new notion of multidimensional order statistics, which may be used to obtain multidimensional outliers. Many of the features revealed using a generalized spatial quantile-based analysis would be missed if the data was shoehorned into a well-known probabilistic configuration.

KW - Brain imaging

KW - Generalized spatial quantile

KW - High dimensional data visualization

KW - Multidimensional coverage sets

KW - Multivariate order statistics

KW - Multivariate quantile

KW - Projection quantile

KW - Spatial quantile

UR - http://www.scopus.com/inward/record.url?scp=79951552574&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79951552574&partnerID=8YFLogxK

U2 - 10.1016/j.jmva.2010.12.002

DO - 10.1016/j.jmva.2010.12.002

M3 - Article

AN - SCOPUS:79951552574

VL - 102

SP - 768

EP - 780

JO - Journal of Multivariate Analysis

JF - Journal of Multivariate Analysis

SN - 0047-259X

IS - 4

ER -