8. MLQuantify Methods

The table below lists all the quantification methods available in the mlquantify library, their references, multiclass support, and type (aggregative, meta, or non-aggregative).

Note

In binary classification problems, methods that do not natively support multiclass classification (marked No in the Multiclass column) remain applicable through standard reduction strategies like one-vs-rest or one-vs-one.

Method	Reference	Multiclass	Type	Module
CC	Forman (2005)	Yes	Aggregative	counting
PCC	Bella et al. (2010)	Yes	Aggregative	counting
GACC	Firat (2016)	Yes	Aggregative	counting
GPACC	Firat (2016)	Yes	Aggregative	counting
TAC	Forman (2005)	No	Aggregative	counting
TX	Forman (2005)	No	Aggregative	counting
TMAX	Forman (2005)	No	Aggregative	counting
T50	Forman (2005)	No	Aggregative	counting
MS	Forman (2006)	No	Aggregative	counting
MS2	Forman (2006)	No	Aggregative	counting
FM	Friedman et al. (2015)	Yes	Aggregative	counting
CDE	Xue & Weiss (2009)	No	Aggregative	likelihood
MLPE	Saerens et al. (2002)	Yes	Aggregative	likelihood
EMQ	Saerens et al. (2002)	Yes	Aggregative	likelihood
DyS	Maletzke et al. (2019)	No	Aggregative	matching
HDy	Gonzalez et al. (2012)	No	Aggregative	matching
SMM	Hassan et al. (2020)	No	Aggregative	matching
SORD	Maletzke et al. (2019)	No	Aggregative	matching
HDx	Gonzalez et al. (2012)	No	Non-aggregative	matching
MMD_RKHS	Iyer et al. (2014)	No	Non-aggregative	matching
KDEyML	Moreo et al. (2025)	Yes	Aggregative	matching
KDEyHD	Moreo et al. (2025)	Yes	Aggregative	matching
KDEyCS	Moreo et al. (2025)	Yes	Aggregative	matching
PWK	Barraquero et al. (2013)	Yes	Aggregative	neighbors
EnsembleQ	Pérez-Gállego et al. (2017) and Pérez-Gállego et al. (2019)	Method dependent	Meta	meta
QuaDapt	Ortega et al. (2025)	Method dependent	Meta	meta
AggregativeBootstrap	Moreo & Salvati (2025)	Method dependent	Meta	meta
QuaNet	Esuli et al. (2018)	Yes	Neural	neural