Cookközépérték

Cookközépérték is a term used in statistics to measure the influence of a single observation on a statistical model, particularly in the context of linear regression. It quantifies how much the predicted value of the dependent variable changes for a unit change in the independent variable when that specific observation is included in the model compared to when it is excluded. A high Cook's distance value indicates that the observation has a significant impact on the regression coefficients and, consequently, on the overall model.

The calculation of Cook's distance involves several steps. It is derived from the residuals of the regression and the leverage of the observation. Leverage measures how far an observation's independent variable values are from the mean of the independent variables. Observations with high leverage are more likely to influence the model. Cook's distance combines the effect of both residual size and leverage.

Generally, a Cook's distance value greater than 1 is considered a threshold for concern, suggesting that the observation may be an outlier or influential point. However, this threshold is not absolute and can vary depending on the dataset and the specific research question. Identifying observations with high Cook's distances is important for diagnostic purposes in regression analysis. These points may warrant further investigation to understand their origin and decide whether they should be retained, transformed, or removed from the dataset.