What do you look for in residual plots?

medium

Answer

$Residuals vs fitted: no pattern = OK; funnel shape = heteroscedasticity; curvature = missed non-linearity.$
$Q-Q plot: check normality of residuals (banana \to skew, sigmoid \to heavy tails).$
$Residuals vs order (for time series): autocorrelation.$
$\mathrm{residuals} \mid \;\mathrm{vs}$
$Leverage vs residuals: identify influential points (Cook's distance).$
$Standard 4-plot diagnostic every regression should get.$

Check yourself — multiple choice

Never plot
Res vs fitted (patterns / funnel); Q-Q for normality; res vs order for autocorr; leverage vs res + Cook's D for influential points
Only for logistic
Same as boxplot

Residual plots: pattern / heteroscedasticity / normality / influential points.

#regression#assumptions#eda

215 interview questions in this topic.