Discovering potential correlations via hypercontractivity

Hyeji Kim, Weihao Gao, Sreeram Kannan, Sewoong Oh, Pramod Viswanath

Research output: Contribution to journalArticle

Abstract

Discovering a correlation from one variable to another variable is of fundamental scientific and practical interest. While existing correlation measures are suitable for discovering average correlation, they fail to discover hidden or potential correlations. To bridge this gap, (i) we postulate a set of natural axioms that we expect a measure of potential correlation to satisfy; (ii) we show that the rate of information bottleneck, i.e., the hypercontractivity coefficient, satisfies all the proposed axioms; (iii) we provide a novel estimator to estimate the hypercontractivity coefficient from samples; and (iv) we provide numerical experiments demonstrating that this proposed estimator discovers potential correlations among various indicators of WHO datasets, is robust in discovering gene interactions from gene expression time series data, and is statistically more powerful than the estimators for other correlation measures in binary hypothesis testing of canonical examples of potential correlations.

Original languageEnglish (US)
Article number586
JournalEntropy
Volume19
Issue number11
DOIs
StatePublished - Nov 1 2017

Fingerprint

axioms
estimators
gene expression
coefficients
genes
estimates
interactions

Keywords

  • Correlation analysis
  • Hypercontractivity
  • Information bottleneck
  • Potential correlation

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Cite this

Discovering potential correlations via hypercontractivity. / Kim, Hyeji; Gao, Weihao; Kannan, Sreeram; Oh, Sewoong; Viswanath, Pramod.

In: Entropy, Vol. 19, No. 11, 586, 01.11.2017.

Research output: Contribution to journalArticle

Kim, Hyeji ; Gao, Weihao ; Kannan, Sreeram ; Oh, Sewoong ; Viswanath, Pramod. / Discovering potential correlations via hypercontractivity. In: Entropy. 2017 ; Vol. 19, No. 11.
@article{2742631482534b508cc4c9efa6f2aa90,
title = "Discovering potential correlations via hypercontractivity",
abstract = "Discovering a correlation from one variable to another variable is of fundamental scientific and practical interest. While existing correlation measures are suitable for discovering average correlation, they fail to discover hidden or potential correlations. To bridge this gap, (i) we postulate a set of natural axioms that we expect a measure of potential correlation to satisfy; (ii) we show that the rate of information bottleneck, i.e., the hypercontractivity coefficient, satisfies all the proposed axioms; (iii) we provide a novel estimator to estimate the hypercontractivity coefficient from samples; and (iv) we provide numerical experiments demonstrating that this proposed estimator discovers potential correlations among various indicators of WHO datasets, is robust in discovering gene interactions from gene expression time series data, and is statistically more powerful than the estimators for other correlation measures in binary hypothesis testing of canonical examples of potential correlations.",
keywords = "Correlation analysis, Hypercontractivity, Information bottleneck, Potential correlation",
author = "Hyeji Kim and Weihao Gao and Sreeram Kannan and Sewoong Oh and Pramod Viswanath",
year = "2017",
month = "11",
day = "1",
doi = "10.3390/e19110586",
language = "English (US)",
volume = "19",
journal = "Entropy",
issn = "1099-4300",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "11",

}

TY - JOUR

T1 - Discovering potential correlations via hypercontractivity

AU - Kim, Hyeji

AU - Gao, Weihao

AU - Kannan, Sreeram

AU - Oh, Sewoong

AU - Viswanath, Pramod

PY - 2017/11/1

Y1 - 2017/11/1

N2 - Discovering a correlation from one variable to another variable is of fundamental scientific and practical interest. While existing correlation measures are suitable for discovering average correlation, they fail to discover hidden or potential correlations. To bridge this gap, (i) we postulate a set of natural axioms that we expect a measure of potential correlation to satisfy; (ii) we show that the rate of information bottleneck, i.e., the hypercontractivity coefficient, satisfies all the proposed axioms; (iii) we provide a novel estimator to estimate the hypercontractivity coefficient from samples; and (iv) we provide numerical experiments demonstrating that this proposed estimator discovers potential correlations among various indicators of WHO datasets, is robust in discovering gene interactions from gene expression time series data, and is statistically more powerful than the estimators for other correlation measures in binary hypothesis testing of canonical examples of potential correlations.

AB - Discovering a correlation from one variable to another variable is of fundamental scientific and practical interest. While existing correlation measures are suitable for discovering average correlation, they fail to discover hidden or potential correlations. To bridge this gap, (i) we postulate a set of natural axioms that we expect a measure of potential correlation to satisfy; (ii) we show that the rate of information bottleneck, i.e., the hypercontractivity coefficient, satisfies all the proposed axioms; (iii) we provide a novel estimator to estimate the hypercontractivity coefficient from samples; and (iv) we provide numerical experiments demonstrating that this proposed estimator discovers potential correlations among various indicators of WHO datasets, is robust in discovering gene interactions from gene expression time series data, and is statistically more powerful than the estimators for other correlation measures in binary hypothesis testing of canonical examples of potential correlations.

KW - Correlation analysis

KW - Hypercontractivity

KW - Information bottleneck

KW - Potential correlation

UR - http://www.scopus.com/inward/record.url?scp=85034248835&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85034248835&partnerID=8YFLogxK

U2 - 10.3390/e19110586

DO - 10.3390/e19110586

M3 - Article

AN - SCOPUS:85034248835

VL - 19

JO - Entropy

JF - Entropy

SN - 1099-4300

IS - 11

M1 - 586

ER -