> [!tldr] > Scatterplot smoothers refer to a general class of methods that model the relationship $f$ in $Y=f(X)+\epsilon$ that make minimal modeling assumptions (in contrast with methods like linear regression, which makes a rather restricting assumption that $f$ maps $X \mapsto X\beta$). > > Examples include [[Localized Methods|local linear regression]], [[Splines|splines]], [[Splines|[[Basis Expansion and Regularization|wavelet smoothing]].