A statistical technique wherein data are compared to certain pre-defined geometric shapes to produce meaningful descriptive simplifications like “average” and “standard deviation”. When applied to test data, this process is sometimes called “data reduction”, where the statistics are referred to as the “reduced data”.
I have observed that many of the non-cognoscenti think that regression can only be done against straight lines. This turns out not to be the case: the “linear” in “linear regression” refers to the coefficients of the independent variables…which, in a perverse turn, are called the “parameters” of the regression equation, not to be confused with the normal use of “parameter” in the rest of the known universe1.
Footnotes- Actually, that’s a lie: it is absolutely to be confused with normal usage, pretty much every time I turn around.[↩]