Overview

Classification studies how to assign inputs such as images, text, signals, graphs, and tabular records to discrete categories, supporting decisions across science and everyday systems. Methods range from linear and kernel discriminants, trees and ensembles, and deep neural networks to pretrained language and vision models, using supervised, few-shot, semi-supervised, and in-context learning. Current work also addresses class imbalance, label noise, distribution shift, open-set recognition, calibration, abstention, and structured labels through robust losses, uncertainty estimates, retrieval, and hierarchical or multimodal models. A central tension is that high benchmark accuracy can conceal shortcut reliance, poor calibration, subgroup failures, or weak transfer, so evaluation increasingly tests reliability, error costs, and performance under deployment shifts.

Latest papers

451 papers45 papers in the last four weeks, down 30% on the four weeks before. 0.3% of all new papers.

Jun 1, 2019

Linear and Quadratic Discriminant Analysis: Tutorial

This tutorial explains Linear Discriminant Analysis (LDA) and Quadratic Discriminant Analysis (QDA) as two fundamental classification methods in statistical and probabilistic learning. We start with the optimization of decision boundary on which the posteriors are equal. Then, LDA and QDA are derived for binary and multiple classes. The estimation of parameters in LDA and QDA are also covered. Then, we explain how LDA and QDA are related to metric learning, kernel principal component analysis, Mahalanobis distance, logistic regression, Bayes optimal classifier, Gaussian naive Bayes, and likelihood ratio test. We also prove that LDA and Fisher discriminant analysis are equivalent. We finally clarify some of the theoretical concepts with simulations we provide.