Table 5 The accuracy measure of different models for CVD and CPD without the consideration of edge weight in the patient network.

From: A weighted patient network-based framework for predicting chronic diseases using graph neural networks

Features

Method

Accuracy for CVD (%)

Accuracy for CPD (%)

With network features (i.e., degree centrality, eigenvector centrality etc.)

LR

76.63

69.34

SVM

79.89

66.51

RF

86.59

79.25

ANN

83.91

73.11

GCN

87.26

85.85

GAT

89.15

88.21

Without network features

GCN

89.08

83.02

GAT

93.49

89.15

Back to article page