Skip to main content

Table 4 Comparison of methods for selecting negative samples based on MDAID dataset

From: Predicting potential microbe-disease associations based on auto-encoder and graph convolution network

Methods

AUC (%)

AUPR (%)

Random sampling

92.88

92.85

KMeans clustering

97.00

96.90

Gaussian mixture clustering

96.95

96.85

Spectral coclustering

93.71

93.98

Spectral biclustering

94.93

94.96

  1. The bold result indicates the best one in each column