Generative Oversampling with a Contrastive Variational Autoencoder

Generative Oversampling with a Contrastive Variational Autoencoder

Show other versions (1)

© 2019 IEEE. Although oversampling methods are widely used to deal with class imbalance problems, most only utilize observed samples in the minority class and ignore the rich information available in the majority class. In this work, we use an oversampling method that leverages information in both t...

Full description

Bibliographic Details
Main Authors: Dai, Wangzhi (Author), Ng, Kenney (Author), Severson, Kristen (Author), Huang, Wei (Author), Anderson, Fred (Author), Stultz, Collin (Author)
Format: Article
Language:English
Published: Institute of Electrical and Electronics Engineers (IEEE), 2021-11-04T17:00:18Z.
Subjects:
Article
Online Access:Get fulltext
Description
Summary:© 2019 IEEE. Although oversampling methods are widely used to deal with class imbalance problems, most only utilize observed samples in the minority class and ignore the rich information available in the majority class. In this work, we use an oversampling method that leverages information in both the majority and minority classes to mitigate the class imbalance problem. Experimental results on two clinical datasets with highly imbalanced outcomes demonstrate that prediction models can be significantly improved using data obtained from this oversampling method when the number of minority class samples is very small.

Similar Items