> [!tldr] Splitting data into groups according to a rule. The terms "Binning" and "Stratifying" are related, but slightly different terms. The have to do with [[Classifiers|classfification]] of samples (i.e. data) into [[Enumeration|allowed "bins"]] for the purposes of further processing. Binning typically implies *evenly sized bin ranges*, whereas stratification typically implies *evenly sized bin population counts*. Because I don't love that phrasing - take this usage of these terms for example: Nielsen ratings "**bin**" the audiences based on **demographics**, eg "18-49" group. If they wanted a representative sample to test a new show on for the *whole* population, they'd **stratify** their sample size so its proportional on a per-demographic basis to the proportion found in the whole population. [[Histogram]]s often use binning. **** # More ## Source - grad school