Table 1.

Correlations between predicted AD values and true values under different models and different selected miRNA panels.

miRNA	Whole genome miRNA expression (1,605)						Spearman correlation (top 400)
Scores	Accuracy	Precision	Recall	F1_score	AUPR	AUC	Accuracy	Precision	Recall	F1_score	AUPR	AUC
SVC	0.8342	0.8926	0.8952	0.8939	0.9087	0.8126	0.7785	0.8741	0.8364	0.8549	0.8837	0.7339
SVR (linear)	0.7217	0.8526	0.7668	0.8074	0.9054	0.7855	0.8171	0.842	0.9213	0.8799	0.9217	0.8253
SVR (rbf)	0.8851	0.879	0.9854	0.9292	0.9274	0.8514	0.8926	0.8862	0.9863	0.9335	0.9273	0.8564
KNN (regression)	0.8235	0.8297	0.9736	0.8959	0.9015	0.7703	0.8296	0.8342	0.9755	0.8993	0.9195	0.806632
miRNA	3D and expression overlap (214)						3D cluster, DBSCAN (100)
Scores	Accuracy	Precision	Recall	F1_score	AUPR	AUC	Accuracy	Precision	Recall	F1_score	AUPR	AUC
SVC	0.8266	0.8825	0.8972	0.8898	0.9173	0.817	0.8449	0.8685	0.9442	0.9047	0.92	0.8186
SVR (linear)	0.8169	0.8169	0.9637	0.8843	0.9197	0.8215	0.7513	0.7465	0.9943	0.8528	0.9132	0.79
SVR (rbf)	0.8853	0.8797	0.9833	0.9286	0.9272	0.8571	0.8686	0.8692	0.9734	0.9183	0.9286	0.8553
KNN (regression)	0.8411	0.8362	0.9902	0.9067	0.9172	0.8002	0.8426	0.8376	0.9902	0.9075	0.9153	0.8052