Consistent estimation of small masses in feature sampling

Battiston, Marco and Ayed, Fadhel and Camerlenghi, Federico and Favaro, Stefano (2021) Consistent estimation of small masses in feature sampling. Journal of Machine Learning Research, 22 (6). pp. 1-28. ISSN 1532-4435

Text (JMLR-18-534-3)
JMLR_18_534_3.pdf - Accepted Version
Available under License Creative Commons Attribution.
Download (499kB)

Abstract

Consider an (observable) random sample of size n from an infinite population of individuals, each individual being endowed with a finite set of features from a collection of features (Fj)j≥1 with unknown probabilities (pj)j≥1, i.e., pj is the probability that an individual displays feature Fj. Under this feature sampling framework, in recent years there has been a growing interest in estimating the sum of the probability masses pj's of features observed with frequency r≥0 in the sample, here denoted by Mn,r. This is the natural feature sampling counterpart of the classical problem of estimating small probabilities in the species sampling framework, where each individual is endowed with only one feature (or “species"). In this paper we study the problem of consistent estimation of the small mass Mn,r. We first show that there do not exist universally consistent estimators, in the multiplicative sense, of the missing mass Mn,0. Then, we introduce an estimator of Mn,r and identify sufficient conditions under which the estimator is consistent. In particular, we propose a nonparametric estimator M^n,r of Mn,r which has the same analytic form of the celebrated Good--Turing estimator for small probabilities, with the sole difference that the two estimators have different ranges (supports). Then, we show that M^n,r is strongly consistent, in the multiplicative sense, under the assumption that (pj)j≥1 has regularly varying heavy tails.