lowess
LOWESS, short for locally weighted scatterplot smoothing, is a nonparametric regression technique used to reveal smooth trends in scatterplot data. The term LOESS, locally estimated scatterplot smoothing, is often used interchangeably. Both methods aim to fit simple models locally to the data rather than assume a single global function.
The core idea is to estimate the value of y at each x by fitting a weighted
Robustness is commonly added by iterating the procedure with robust weights that reduce the influence of outliers.
Differences and usage: LOWESS typically refers to the linear local fit, while LOESS often implies higher-order