Good-Turing Estimator

A statistical estimation method for estimating the probability of unobserved events, originally proposed by Good and Turing. It adjusts probabilities based on the frequencies of samples occurring r times and is widely used in natural language processing for smoothing and missing-mass estimation.