A Fast and Efficient Estimation of the Parameters of a Model of Accident Frequencies via an MM Algorithm

Table 2

Results for Scenario 2 ( and ). Values in brackets are standard deviations.

		MM	SqS3	NR	BFGS
	Convergence proportion (%)	100	100	99.6	50.4
	Iterations	29.3 (4.2)	7.8 (0.8)	8.1 (3)	14 (0.1)
	CPU time (secs)	0.007	0.003	0.015	0.283
	Time ratio	1.00	0.47	2.14	39.93
	Log-likelihood	-391.24	-391.24	-391.21	-391.6
	MSE
	Convergence proportion (%)	100	100	99.7	53.3
	Iterations	38.2 (4.1)	9.1 (0.8)	8.4 (2.9)	18.3 (3.4)
	CPU time (secs)	0.009	0.004	0.016	0.533
	Time ratio	1.00	0.43	1.80	59.55
	Log-likelihood	-39181.57	-39181.57	-39181.5	-39942.08
	MSE