A robust approach to model-based classification based on trimming and constraints. Semi-supervised learning in presence of outliers and label noise

DOI10.1007/S11634-019-00371-WzbMATH Open1474.62215arXiv1904.06136OpenAlexW3121868169WikidataQ127361911 ScholiaQ127361911MaRDI QIDQ2201323FDOQ2201323

Thomas Brendan Murphy, Francesca Greselin, Andrea Cappozzo

Publication date: 29 September 2020

Published in: Advances in Data Analysis and Classification. ADAC (Search for Journal in Brave)

Abstract: In a standard classification framework a set of trustworthy learning data are employed to build a decision rule, with the final aim of classifying unlabelled units belonging to the test set. Therefore, unreliable labelled observations, namely outliers and data with incorrect labels, can strongly undermine the classifier performance, especially if the training size is small. The present work introduces a robust modification to the Model-Based Classification framework, employing impartial trimming and constraints on the ratio between the maximum and the minimum eigenvalue of the group scatter matrices. The proposed method effectively handles noise presence in both response and exploratory variables, providing reliable classification even when dealing with contaminated datasets. A robust information criterion is proposed for model selection. Experiments on real and simulated data, artificially adulterated, are provided to underline the benefits of the proposed method.

Full work available at URL: https://arxiv.org/abs/1904.06136

Recommendations

• Anomaly and novelty detection for robust semi-supervised learning
• Robust supervised classification with mixture models: learning from data with uncertain labels
• Robust variable selection for model-based learning in presence of adulteration
• Learning kernel logistic regression in the presence of class label noise
• A robust discriminant approach and its application

zbMATH Keywords

robust estimationmodel-based classificationoutliers detectionlabel noiseimpartial trimmingeigenvalues restrictions

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Learning and adaptive systems in artificial intelligence (68T05) Robustness and adaptive procedures (parametric inference) (62F35) Compound decision problems in statistical decision theory (62C25)

Cites Work

• A decision-theoretic generalization of on-line learning and an application to boosting
• Estimating common principal components in high dimensions
• Robust fitting of mixtures using the trimmed likelihood estimator
• Robust inference for parsimonious model-based clustering
• Estimating the dimension of a model
• Title not available (Why is that?)
• Title not available (Why is that?)
• Model-Based Clustering, Discriminant Analysis, and Density Estimation
• The distribution of the likelihood ratio for mixtures of densities from the one-parameter exponential family
• Support-vector networks
• Model-Based Gaussian and Non-Gaussian Clustering
• Title not available (Why is that?)
• Exploring the number of groups in robust model-based clustering
• A general trimming approach to robust cluster analysis
• Trimmed \(k\)-means: An attempt to robustify quantizers
• The EM Algorithm and Extensions, 2E
• High-Breakdown Linear Discriminant Analysis
• Regularized Gaussian Discriminant Analysis Through Eigenvalue Decomposition
• Title not available (Why is that?)
• Using Unlabelled Data to Update Classification Rules with Applications in Food Authenticity Studies
• A likelihood-based constrained algorithm for multivariate normal mixture models
• A fast algorithm for robust constrained clustering
• Finding the Number of Normal Groups in Model-Based Clustering via Constrained Likelihoods
• Avoiding spurious local maximizers in mixture modeling
• A constrained robust proposal for mixture modeling avoiding spurious solutions
• Best approximations to random variables based on trimming procedures
• A review of robust clustering methods
• Title not available (Why is that?)
• Robust supervised classification with mixture models: learning from data with uncertain labels
• Multivariate Clustering Procedures with Variable Metrics
• Class noise vs. attribute noise: A quantitative study of their impacts
• Noise modelling and evaluating learning from examples
• On the breakdown point of multivariate location estimators based on trimming procedures
• A reweighting approach to robust clustering
• The joint role of trimming and constraints in robust estimation for mixtures of Gaussian factor analyzers
• Density-based silhouette diagnostics for clustering methods

Cited In (6)

• Robust variable selection for model-based learning in presence of adulteration
• Distance-based directional depth classifiers: a robustness study
• Harmless label noise and informative soft-labels in supervised classification
• Robust image classification
• Consistency factor for the MCD estimator at the Student-\(t\) distribution
• Robust supervised classification with mixture models: learning from data with uncertain labels

Uses Software

• mclust
• R
• CRAN
• AdaBoost.MH
• TCLUST
• upclass
• mtclust

This page was built for publication: A robust approach to model-based classification based on trimming and constraints. Semi-supervised learning in presence of outliers and label noise

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2201323)