ImputeINR: Time Series Imputation via Implicit Neural Representations for Disease Diagnosis with Missing Data

TL;DR

ImputeINR introduces a novel neural representation-based method for time series imputation, effectively handling sparse healthcare data with high missing ratios and improving disease diagnosis accuracy.

Contribution

The paper presents ImputeINR, a new approach using implicit neural representations for continuous, sampling-frequency-independent time series imputation, especially effective for highly sparse data.

Findings

Outperforms existing methods on eight datasets with high missing ratios.

Generates fine-grained imputations even with extremely sparse observed data.

Enhances downstream disease diagnosis performance when applied to healthcare data.

Abstract

Healthcare data frequently contain a substantial proportion of missing values, necessitating effective time series imputation to support downstream disease diagnosis tasks. However, existing imputation methods focus on discrete data points and are unable to effectively model sparse data, resulting in particularly poor performance for imputing substantial missing values. In this paper, we propose a novel approach, ImputeINR, for time series imputation by employing implicit neural representations (INR) to learn continuous functions for time series. ImputeINR leverages the merits of INR in that the continuous functions are not coupled to sampling frequency and have infinite sampling frequency, allowing ImputeINR to generate fine-grained imputations even on extremely sparse observed values. Extensive experiments conducted on eight datasets with five ratios of masked values show the superior imputation performance of ImputeINR, especially for high missing ratios in time series data. Furthermore, we validate that applying ImputeINR to impute missing values in healthcare data enhances the performance of downstream disease diagnosis tasks. Codes are available.