distance covariance
E472795
measure of dependence
multivariate dependence measure
nonparametric dependence measure
statistical measure
Distance covariance is a statistical measure that quantifies dependence between random variables, capable of detecting both linear and nonlinear associations.
Statements (46)
| Predicate | Object |
|---|---|
| instanceOf |
measure of dependence
ⓘ
multivariate dependence measure ⓘ nonparametric dependence measure ⓘ statistical measure ⓘ |
| advantageOver | Pearson correlation in detecting nonlinear dependence ⓘ |
| belongsTo |
dependence modeling
ⓘ
multivariate statistics ⓘ nonparametric statistics ⓘ |
| canDetect |
linear dependence
ⓘ
nonlinear dependence ⓘ |
| comparedWith |
Pearson correlation
ⓘ
mutual information ⓘ |
| dependsOn | Euclidean distances between sample points ⓘ |
| equalsZeroFor | independent random variables ⓘ |
| generalizes | classical covariance in detecting dependence ⓘ |
| greaterThanZeroFor | dependent random variables ⓘ |
| hasExtension |
conditional distance covariance
ⓘ
distance covariance for time series ⓘ partial distance covariance ⓘ |
| hasNormalizedForm | distance correlation ⓘ |
| hasProperty | characterizes independence in Euclidean spaces with finite first moments ⓘ |
| implementedIn |
Python libraries
ⓘ
R packages ⓘ |
| introducedBy |
Gábor J. Székely
NERFINISHED
ⓘ
Maria L. Rizzo NERFINISHED ⓘ |
| introducedIn | 2007 ⓘ |
| invariantUnder |
orthogonal transformations of the data
ⓘ
translations of the data ⓘ |
| isBasedOn | pairwise distances between observations ⓘ |
| isDefinedFor |
multivariate random variables
ⓘ
random vectors in arbitrary dimensions ⓘ univariate random variables ⓘ |
| isEstimatedBy | sample distance covariance ⓘ |
| isNonNegative | true ⓘ |
| isZeroIfAndOnlyIf | random variables are independent ⓘ |
| publishedIn | Annals of Statistics NERFINISHED ⓘ |
| quantifies | dependence between random variables ⓘ |
| relatedTo |
Brownian covariance
ⓘ
distance correlation ⓘ |
| requires |
choice of metric space for the variables
ⓘ
finite first moments of the random variables ⓘ |
| symmetricIn | its two arguments ⓘ |
| usedFor |
feature screening
ⓘ
measuring association in high dimensions ⓘ testing independence ⓘ variable selection ⓘ |
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.