Iterative Reclassification Procedure for Constructing An Asymptotically Optimal Rule of Allocation in Discriminant-Analysis

By Geoffrey J. McLachlan. 1975

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistical Learning Theory

Citations in the corpus

FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (2020, background): “Pseudo-labeling leverages the idea of using the model itself to obtain artificial labels for unlabeled data [32, 47].”
Semi-supervised Learning with Ladder Networks (2015, background): “Early works in semi-supervised learning (McLachlan, 1975; Titterington et al. , 1985) proposed an approach where inputs x are ﬁrst assigned to clusters, and each cluster has its class label.”
ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring (2019, methodology): “Related to this idea are “self-training” methods (McLachlan, 1975; Rosenberg et al., 2005) such as Pseudo-Label (Lee, 2013) that use the predicted class on an unlabeled input as a hard target for the same input, which implicitly minimizes the entropy of the prediction.”
Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper (1977, methodology): “McClachlan (1975) has given a procedure for using unclassified observations in discriminant analysis to help estimate the discriminant function.”
A Review of Domain Adaptation without Target Labels (2019, background): “This is known as iterative self-labeling or self-learning [182].”
Enhancing Person Re-identification in a Self-Trained Subspace (2017, background): “[43] is one of the earliest works that applies this strategy to design an iterative reclassification procedure.”
FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling (2021, methodology): “With the introduction of pseudo labeling techniques [5, 7], the consistency regularization is converted to an entropy minimization process [16], which is more suitable for the classiﬁcation task.”
Projected estimators for robust semi-supervised classification (2016, background): “Early work on semi-supervised learning dealt with the missing labels through the use of Expectation Maximization in generative models or closely related self-learning (McLachlan, 1975).”
Better Self-training for Image Classification through Self-supervision (2021, background): “Semi-supervised learning approaches have been developed since the 1970s [14], and self-training, in particular, has continued to attract interest.”
Using unlabelled data to update classification rules with applications in food authenticity studies (2006, methodology): “that have appeared in Ganesalingam and McLachlan (1978) and O’Neill (1978), among others; these related results are briefly reviewed in Section 4.”
Dash: Semi-Supervised Learning with Dynamic Thresholding (2021, methodology): “selftraining (McLachlan, 1975; Yarowsky, 1995; Rosenberg et al., 2005; Sajjadi et al., 2016; Laine & Aila, 2017; Xie et al., 2020b)) uses the prediction model itself to generate pseudo labels for unlabeled images.”
Semi-supervised learning by disagreement (2010, unknown): “In fact, some straightforward use of unlabeled examples appeared even earlier [40] [50] [52] [53] [57].”
FreeMatch: Self-adaptive Thresholding for Semi-supervised Learning (2022, background): “Pseudo labeling [11, 12, 13, 14] and consistency regularization [15, 16, 17] are two popular paradigms designed for modern SSL.”
A Simple Semi-Supervised Learning Framework for Object Detection (2020, methodology): “To this end, we propose a SSL framework for object detection that combines self-training (via pseudo label) [47, 34] and consistency regularization based on the strong data augmentations [6, 7, 70].”
Contrastive Pessimistic Likelihood Estimation for Semi-Supervised Classification (2015, background): “McLachlan [33], in 1975, was probably the ﬁrst to apply this procedure and indeed suggested it as a computationally more tractable alternative to the one in [32].”
Adaptive computation and machine learning (1998, methodology): “McLachlan (McLachlan, 1975) gives a practical algorithm for this case which is essentially a “hard” version of EM, i.”
Semi-Supervised Text Classification Using EM (2006, methodology): “Similar iterative algorithms for building maximum likelihood classifiers from labeled and unlabeled data with an explicit generative model followed, primarily for mixtures of normal distributions (McLachlan [1975], Titterington [1976]). Dempster et al. [1977] presented the theory of the EM framework”