Table 3.

Model importance factors (random forest) for follicle diameter in determining number of oocytes and number of mature oocytes.

Follicle diameter (mm)	hCG	GnRH agonist (GnRHa)	Kisspeptin	hCG	GnRHa	Kisspeptin
	Oocytes			Mature oocytes
	Model importance factor	Model importance factor	Model importance factor	Model importance factor	Model importance factor	Model importance factor
8	18.9	1.3	177.3	25.8	0.7	241.4
9	42.2	0.2	102.8	48.0	0.3	107.0
10	41.4	19.8	66.1	34.7	23.3	77.5
11	38.7	54.8	103.4	35.9	53.2	88.5
12	104.9	158.9	88.8	113.5	150.4	84.9
13	148.9	109.8	172.6	197.8	66.1	143.7
14	296.3	340.7	197.1	239.8	340.9	110.7
15	207.8	335.9	58.8	215.7	360.0	77.6
16	190.7	172.8	222.6	187.2	154.3	312.0
17	255.8	287.5	220.3	256.3	312.8	214.9
18	189.5	132.4	107.5	186.5	124.4	77.0
19	111.2	109.4	41.2	102.4	151.7	47.6
20	71.6	46.8	51.9	92.3	36.8	42.9
21	28.2	20.0	29.5	26.3	16.6	30.9
22	30.6	7.8	19.4	22.5	7.5	32.7
23	12.0	1.9	23.5	7.1	1.0	19.7
24	11.3	0.0	22.4	8.3	0.0	18.8
25	0.0	0.0	15.4	0.0	0.0	18.7

Model importance factors are the increases in node purity given by each follicle diameter averaged over 5,000 classification trees, each derived using bootstrap samples from the study data. Values over 100 (in bold) denote follicle diameters that are consistently important as predictors across multiple models, and hence strengthen the P-value results given in Table 2 which could result from a chance configuration of the study data.