Random Forests in Count Data Modelling: An Analysis of the Influence of Data Features and Overdispersion on Regression Performance

Table 2

Effect of predictor types and dispersion amplitude on the number of variables randomly selected at each split.

Data types	Variance-to-mean relationship	Best mtry	N = 50 (%)						N = 250 (%)						N = 1250 (%)
			Categorical	Linear	2	81	89	90	56	58	65	98	97	99	80	84	86	100	100	100	91	92	94
						9	3	6	27	27	30	1	3	1	17	15	14	0	0	0	9	8	6
	8	7			4	9	11	5	0	0	0	0	0	0	0	0	0	0	0	0
	2	1			0	8	4	0	1	0	0	3	1	0	0	0	0	0	0	0
Quadratic	2	89		88	88	62	74	74	98	99	99	81	85	92	100	100	100	97	99	98
		7		8	4	28	16	18	2	0	1	18	15	8	0	0	0	3	1	2
		2		3	1	4	7	6	0	1	0	1	0	0	0	0	0	0	0	0
		2		1	7	6	3	2	0	0	0	0	0	0	0	0	0	0	0	0
25% of predictors are quantitative	Linear	2	0	0	100	0	100	100	100	100	100	100	0	100	100	100	100	100	100	100
			0	0	0	0	0	0	0	0	0	0	100	0	0	0	0	0	0	0
			100	0	0	100	0	0	0	0	0	0	0	0	0	0	0	0	0	0
			0	100	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
	Quadratic	2	100	100	100	100	100	0	100	100	100	100	0	100	100	100	100	100	100	100
			0	0	0	0	0	0	0	0	0	0	100	0	0	0	0	0	0	0
			0	0	0	0	0	100	0	0	0	0	0	0	0	0	0	0	0	0
50% of predictors are quantitative	Linear	2	100	100	100	0	0	0	100	100	0	0	100	100	100	100	100	100	100	100
			0	0	0	0	0	100	0	0	100	100	0	0	0	0	0	0	0	0
			0	0	0	100	0	0	0	0	0	0	0	0	0	0	0	0	0	0
			0	0	0	0	100	0	0	0	0	0	0	0	0	0	0	0	0	0
	Quadratic	2	100	100	100	100	100	100	100	100	100	100	100	100	100	100	100	100	100	100
75% of predictors are quantitative	Linear	2	100	100	100	0	100	100	100	100	100	100	0	100	100	100	100	100	100	100
			0	0	0	100	0	0	0	0	0	0	100	0	0	0	0	0	0	0
	Quadratic	2	100	0	100	0	100	100	100	100	100	100	100	100	100	100	100	100	100	100
			0	100	0	100	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Quantitative	Linear	2	100	0	100	100	100	100	0	100	100	0	100	100	100	100	100	100	100	100
			0	100	0	0	0	0	100	0	0	100	0	0	0	0	0	0	0	0
	Quadratic	2	100	100	0	100	100	100	100	100	100	100	100	100	100	100	100	100	100	100
			0	0	100	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0