Learning with Positive and Unlabeled Examples Using Weighted Logistic Regression

By Wee Sun Lee, B. Liu. International Conference on Machine Learning, 2003

416 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/41d9b4104b6d8cb8c135560b5f775bc8cd7a720d

How researchers see this paper

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

Citations in the corpus

A bagging SVM to learn from positive and unlabeled examples (2010, methodology): “This approach has been studied, with different weighting schemes, using a logistic regression or a SVM as binary classifier (Liu et al., 2003; Lee and Liu, 2003; Elkan and Noto, 2008).”
One-class classification: taxonomy of study and review of techniques (2013, unknown): “In this section, we review some major OCC algorithms that are based on one-class ensembles, neural networks, decision trees, nearest neighbors, Bayesian classifiers and other methods.”
Social Media Text Classification under Negative Covariate Shift (2015, background): “Our work is also related to learning from positive and unlabeled examples, also known as PU learning (Denis, 1998; Yu et al. 2002; Liu et al. 2003; Lee and Liu, 2003; Elkan and Noto, 2008; Li et al. 2010).”
Learning from positive and unlabeled data: a survey (2018, methodology): “ata, but something similar can be. Note that the F 1 score is high when both precision and recall are high. The following performance criterion has the same property and can be estimated from PU data [52]: pr Pr(y = 1) = pr2 rPr(y = 1) = Pr(y = 1jy^ = 1)r2 Pr(^y = 1;y = 1) = r2 Pr(y^ = 1) : (10) 4.2”
Positive-Unlabeled Learning with Non-Negative Risk Estimator (2017, background): “The second (e.g., Lee & Liu, 2003; Liu et al., 2003) regards all U data as N data but they are weighted smaller.”
Identifying Where to Focus in Reading Comprehension for Neural Question Generation (2017, methodology): “For the related important sentence detection task in text summarization, Yang et al. (2017) therefore propose a two-stage approach (Lee and Liu, 2003; Elkan and Noto, 2008) to augment the set of known summaryworthy sentences.”
Unsupervised Representation Learning for Time Series with Temporal Neighborhood Coding (2021, methodology): “…PU learning methods fall under two categories based on how they handle the unlabeled data: 1) methods that identify negative samples from the unlabeled cohort (Li & Liu, 2003); 2) methods that treat the unlabeled data as negative samples with smaller weights (Lee & Liu, 2003; Elkan & Noto, 2008).”
Learning classifiers from only positive and unlabeled data (2008, methodology): “This approach is used for example by [8, 12].”
Semi-Supervised Novelty Detection (2010, methodology): “We present a hybrid approach that automatically reverts to the inductive approach when π = 0, while preserving the benefits of the NP reduction whenπ > 0.”
Outlier Analysis (2013, background): “A less common approach is to assign weights to the unlabeled training examples [293, 306].”
Introduction to Semi-Supervised Learning (2009, background): “• learning from positive and unlabeled data,when there is no negative labeled data [61,114,109];”
Building text classifiers using positive and unlabeled examples (2003, methodology): “In [16], it is shown that pr/Pr[Y = 1] = r(2)/Pr[f(X) = 1], where Pr[f(X) = 1] is the probability that a document is classified as positive.”
A Survey of Recent Trends in One Class Classification (2009, background): “The last category has recieved much research interest among the text/document classification community [8][9][10] that will be discussed in detail below in Section 3.”
Learning with Confident Examples: Rank Pruning for Robust Classification with Noisy Labels (2017, methodology): “Early approaches to PU learning modified the loss functions via weighted logistic regression (Lee & Liu, 2003) and biased SVM (Liu et al., 2003) to penalize more when positive examples are predicted incorrectly.”