Autoencoder-based Attribute Noise Handling Method for Medical Data

TL;DR

This paper introduces an autoencoder-based preprocessing technique to effectively correct attribute noise in medical tabular data, significantly improving data quality before analysis.

Contribution

It presents the first method specifically designed to handle attribute noise in tabular medical datasets using autoencoders, outperforming existing imputation and noise correction techniques.

Findings

Outperforms state-of-the-art imputation methods

Effective in correcting mixed-type attribute noise

Enhances data quality for medical datasets

Abstract

Medical datasets are particularly subject to attribute noise, that is, missing and erroneous values. Attribute noise is known to be largely detrimental to learning performances. To maximize future learning performances it is primordial to deal with attribute noise before any inference. We propose a simple autoencoder-based preprocessing method that can correct mixed-type tabular data corrupted by attribute noise. No other method currently exists to handle attribute noise in tabular data. We experimentally demonstrate that our method outperforms both state-of-the-art imputation methods and noise correction methods on several real-world medical datasets.