Indirect estimation of pediatric reference interval via density graph deep embedded clustering

Jianguo Zheng, Yongqiang Tang, XiaoXia Peng, Jun Zhao, Rui Chen, Ruohua Yan, Yaguang Peng, Wensheng Zhang

Abstract: Highlights•This study introduces deep graph clustering methods into indirect estimation of pediatric reference intervals.•We propose a Density Graph Deep Embedded Clustering (DGDEC) algorithm, which incorporates a density feature extractor to enhance sample representation and provides additional perspectives for distinguishing different levels of health status among populations.•We utilize an elaborate graph autoencoder to exploit structural correlation among patient samples. Through this way, a more ideal clustering performance can be achieved to segment patient samples into different collections of health conditions, providing a novel technical implement for pediatric RI estimation.•Experimental results on real-world dataset demonstrate that the proposed method can achieve favorable performance and robustness compared with the current state-of-the-art indirect estimation methods.