Predicting Metabolite-Disease Associations Based on Linear Neighborhood Similarity with Improved Bipartite Network Projection Algorithm

Table 2

Parameter analysis for kd.

kd	10	20	30	40	50
AUCs	0.8553	0.8542	0.8589	0.8579	0.8563

kd is changed, and km is set 50 under 10-fold cross validation.