Cook’s Distance

Formula

The distance 𝐷_𝑖 for data point 𝑖 is defined as:

• $D_i=\frac{{\sum }_{j=1}^n({\hat{y}}_j-{\hat{y}}_{j(i)}{)}^2}{p{\widehat{𝜎}}^2}$

where:

• 𝑦̂_𝑗(𝑖) - is the fitted value for data point 𝑗 with the 𝑖^th observation removed