A bagging SVM to learn from positive and unlabeled examples

By Fantine Mordelet, Jean-Philippe Vert. Pattern Recognition Letters, 2010

326 citations recorded by the source corpus.

We consider the problem of learning a binary classifier from a training set of positive and unlabeled examples, both in the inductive and in the transductive setting. This problem, often referred to as PU learning, differs from the standard supervised classification problem by the lack of negative examples in the training set. It corresponds to an ubiquitous situation in many applications such as information retrieval or gene ranking, when we have identified a set of data of interest sharing a particular property, and we wish to automatically retrieve additional data sharing the same property among a large and easily available pool of unlabeled data. We propose a new method for PU learning with a conceptually simple implementation based on bootstrap aggregating (bagging) techniques: the algorithm iteratively trains many binary classifiers to discriminate the known positive examples from random subsamples of the unlabeled set, and averages their predictions. We show theoretically and experimentally that the method can match and even outperform the performance of state-of-the-art methods for PU learning, particularly when the number of positive examples is limited and the fraction of negatives among the unlabeled examples is small. The proposed method can also run considerably faster than state-of-the-art methods, particularly when the set of unlabeled examples is large.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/36adc7425452dea712e1387db2aceabceb092111

How researchers see this paper

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

In Artificial Intelligence

Citations in the corpus

Learning from positive and unlabeled data: a survey (2018, methodology): “Several PU methods were developed to this end: ProDiGe is a method based on bagging SVMs [70,72], PUDI is also a weighted SVM method, but they have different weights for four identified groups of unlabeled examples: reliable negative, likely positive, likely negative and weakly negative [110],”
Learning with Confident Examples: Rank Pruning for Robust Classification with Noisy Labels (2017, methodology): “Bagging SVM (Mordelet & Vert, 2014) and RESVM (Claesen et al., 2015) extended biased SVM to instead use an ensemble of classifiers trained by resampling U (and P for RESVM) to improve robustness (Breiman, 1996).”
Sensing spatial distribution of urban land use by integrating points-of-interest and Google Word2Vec model (2017, background): “2011, Huang and Zhang 2013, Mordelet and Vert 2014).”
Low-Resource Corpus Filtering Using Multilingual Sentence Embeddings (2019, methodology): “(Mordelet and Vert, 2014), which allows us to obtain classifiers without having to curate a dataset of explicit positive and negatives.”
Fast Threshold Tests for Detecting Discrimination (2017, background): “The setup thus bears resemblance to learning classiﬁers from only positive examples (Elkan & Noto, 2008; Mordelet & Vert, 2014; du Plessis et al., 2014).”