AdaMedGraph: Adaboosting Graph Neural Networks for Personalized Medicine

TL;DR

AdaMedGraph introduces an automated method to select features for constructing multiple patient similarity graphs, enhancing graph neural network performance in personalized medicine prediction tasks.

Contribution

It proposes a novel algorithm that automatically identifies important features for graph construction and integrates them into GNNs using adaptive boosting, reducing reliance on human expertise.

Findings

Outperforms existing methods on real-world medical datasets

Automatically selects relevant features for graph construction

Enhances prediction accuracy in personalized medicine

Abstract

Precision medicine tailored to individual patients has gained significant attention in recent times. Machine learning techniques are now employed to process personalized data from various sources, including images, genetics, and assessments. These techniques have demonstrated good outcomes in many clinical prediction tasks. Notably, the approach of constructing graphs by linking similar patients and then applying graph neural networks (GNNs) stands out, because related information from analogous patients are aggregated and considered for prediction. However, selecting the appropriate edge feature to define patient similarity and construct the graph is challenging, given that each patient is depicted by high-dimensional features from diverse sources. Previous studies rely on human expertise to select the edge feature, which is neither scalable nor efficient in pinpointing crucial edge features for complex diseases. In this paper, we propose a novel algorithm named \ours, which can automatically select important features to construct multiple patient similarity graphs, and train GNNs based on these graphs as weak learners in adaptive boosting. \ours{} is evaluated on two real-world medical scenarios and shows superiors performance.