Learning from imbalanced data sets with weighted cross-entropy function
Carregando...
Data
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Minas Gerais
Descrição
Tipo
Artigo de periódico
Título alternativo
Primeiro orientador
Membros da banca
Resumo
This paper presents a novel approach to deal with the imbalanced data set problem in neural networks by incorporating prior probabilities into a cost-sensitive cross-entropy error function. Several classical benchmarks were tested for performance evaluation using different metrics, namely G-Mean, area under the ROC curve (AUC), adjusted G-Mean, Accuracy, True Positive Rate, True Negative Rate and F1-score. The obtained results were compared to well-known algorithms and showed the effectiveness and robustness of the proposed approach, which results in well-balanced classifiers given different imbalance scenarios.
Abstract
Assunto
Redes neurais (Computação)
Palavras-chave
The number of samples commonly differs from one class to another in classification problems. This problem, known as the imbalanced data set problem, arises in most real-world applications. The point is that most current inductive learning principles resides on a sum of squared errors that do not take priors into account, which generally results in a classification bias towards the majority class.
Citação
Curso
Endereço externo
https://link.springer.com/article/10.1007/s11063-018-09977-1