Gabriel graph transductive approach to dataset shift
Carregando...
Data
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Minas Gerais
Descrição
Tipo
Artigo de evento
Título alternativo
Primeiro orientador
Membros da banca
Resumo
It is not uncommon for data obtained from systems to change after the model is learned. These occurrences are named dataset shifts and to deal with them models with the ability to adapt to data changes must be used. A strategy that can be easily integrated to other classifiers is proposed. It creates a geometrical representation of data that extracts information from both labelled and unlabelled data. Then data entropy and Jensen-Shannon dissimilarity tests are used during the model selection to handle cases where data shift. Results have shown that the proposed method is promising because of its simple integration with state of art classifiers and its performance in enhancing said classifiers accuracy in the studied cases.
Abstract
Assunto
Modelos matemáticos
Palavras-chave
Data models , Entropy , Gaussian mixture model , Labeling , Adaptation models , Mathematical model, Dataset Shift , Transduction Approach , Gabriel Graph , Unlabeled Data , Training Data , Random Variables , Support Vector Machine , Covariance Matrix , Probability Density Function , Mixture Model , Data Clustering , Radial Basis Function , Spatial Clustering , Gaussian Mixture Model , Binary Classification Problem , Labeling Process , Cluster Labels , Delaunay Triangulation , State Of The Art Methods , Jensen-Shannon Divergence , Cluster Boundaries , Cluster Aggregation , Improvement In Classification , Multilayer Perception
Citação
Departamento
Curso
Endereço externo
https://ieeexplore.ieee.org/abstract/document/8820327