Conditional Variational Autoencoder-Based Sampling

Files

Access Instruction 743.pdf (45.1 KB)

Date

2023

Authors

Journal Title

Journal ISSN

Volume Title

Publisher

Springer Science and Business Media Deutschland GmbH

Abstract

Imbalanced data distribution implies an uneven distribution of class labels in data which can lead to classification bias in machine learning models. The present paper proposes an autoencoder-based sampling approach to balance the data. Concretely, the proposed method utilizes a conditional variational autoencoder (VAE) to learn the latent variables underpinning the distribution of minority labels. Then, the trained encoder is employed to produce new minority samples to equalize the sample distribution. The results of numerical experiments reveal the potency of the suggested technique on several datasets. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Description

Keywords

Autoencoder, Imbalanced data, Sampling

Citation

Kamalov, F., Ali-Gombe, A., Moussa, S. (2023). Conditional Variational Autoencoder-Based Sampling. In: Fong, S., Dey, N., Joshi, A. (eds) ICT Analysis and Applications. Lecture Notes in Networks and Systems, 517, pp. 661 - 669. Springer, Singapore. https://doi.org/10.1007/978-981-19-5224-1_66

URI

https://hdl.handle.net/20.500.12519/743

DOI

Collections

Department of Electrical Engineering