Statistical Analysis Stack Exchange
Statistical Analysis Stack Exchange

Questions tagged [cooks-distance]

A measure of the influence of a single observation in regression modelling.

Cook's distance is a measure of the influence a single observation has. It is sometimes referred to as Cook's deleted residual statistic. It is a function of the distance which the parameter estimates would move if the observation were deleted. Further details can be obtained from the Wikipedia entry or from Cook's original article.

68 questions
3
votes
1 answer

Difference between DFBETA with DFFITS / Cook’s distance

A description of DFBETA is on the Wikipedia page: https://en.wikipedia.org/wiki/Influential_observation along with DFFITS and Cook’s distance. Cook’s distance and DFFITS are conceptually identical and there is a formula to convert one value to the…
Single Malt
  • 564
  • 1
  • 8
  • 21
2
votes
0 answers

Cook's distance for other then linear regression methods

I am curious when I can use cook's distance. I read here that it is only used for linear regression, but then on another site that it can be used for regression trees. So when is it appropriate to use Cook's distance. Can we use it for RF or…
Klemz
  • 43
1
vote
0 answers

Cook's distance threshold in R

Short question: Which method is used in R to determine the threshold of the Cook's distance? For example in the demand '''ols_plot_cooksd_bar''' there is drawn a threshold line.
Marie Vermeiren
  • 11