Table 5. Prediction performance for 3-month treatment outcomes. Parentheses denote standard deviations in cross validation.

Target	Model	Validation BS	Testing BS
<6.5%	RF	0.2294 (0.0256)	0.1634
	GBT	0.1684 (0.0198)	0.1430
	TPGE	0.1315 (0.0025)	0.1389
<7.0%	RF	0.1949 (0.0237)	0.1939
	GBT	0.1961 (0.0243)	0.1616
	TPGE	0.1437 (0.0032)	0.1576
<7.5%	RF	0.2741 (0.0569)	0.1682
	GBT	0.1727 (0.0067)	0.1448
	TPGE	0.1319 (0.0027)	0.1405
<8.0%	RF	0.2197 (0.0828)	0.1307
	GBT	0.1317 (0.0102)	0.1121
	TPGE	0.1059 (0.0022)	0.1098
<8.5%	RF	0.0944 (0.0040)	0.1030
	GBT	0.0945 (0.0040)	0.0854
	TPGE	0.0809 (0.0018)	0.0850
Average	RF	0.2025 (0.0386)	0.1518
	GBT	0.1527 (0.0130)	0.1294
	TPGE	0.1188 (0.0025)	0.1264

Abbreviations: BS, Brier Score; GBT, Gradient Boosting Tree; RF, Random Forest; TPGE, Treatment Pathway Graph-based Estimation.