New Cure Rate Survival Models Generated by Poisson Distribution and Different Regression Structures with Applications to Cancer Data Set

Table 8

Model comparison using the simulated average of AIC and BIC and the percentage of the lowest AIC and BIC corresponding to the link functions in the simulation studies with covariates.

True link is the logit
Model		Information criteria	Logit	Cauchit	Aranda-Ordaz	Probit	Log-log	Skewed logit	Skewed probit
PGcr	150	Average of AIC (% lowest AIC)	−433.57 (%17.48)	−438.88 (%17.21)	−420.12 (%11.74)	−434.56 (%13.66)	−460.97 (%23.22)	−386.17 (%6.01)	−420.85 (%10.65)
		Average of BIC (% lowest BIC)	−406.47 (%17.48)	−411.78 (%17.48)	−390.01 (%11.47)	−407.46 (%13.93)	−433.88 (%23.77)	−356.60 (%5.73)	−390.75 (%10.10)
	300	Average of AIC (% lowest AIC)	−856.61 (%15.64)	−808.33 (%6.44)	−853.4 (%17.48)	−860.88 (%17.48)	−873.52 (%17.48)	−823.93 (%10.42)	−858.05 (%15.03)
		Average of BIC (% lowest BIC)	−823.27 (%15.64)	−775.00 (%6.44)	−816.36 (%16.25)	−827.55 (%18.71)	−840.19 (%17.79)	−786.89 (%10.42)	−821.02 (%14.72)
PGEcr	150	Average of AIC (% lowest AIC)	296.00 (%16.6)	297.54 (%14.33)	299.80 (%10.56)	297.03 (%13.77)	296.93 (%16.41)	296.95 (%15.66)	298.82 (%12.64)
		Average of BIC (% lowest BIC)	320.09 (%16.98)	321.62 (%16.60)	326.90 (%8.49)	321.11 (%16.22)	321.01 (%17.54)	324.04 (%14.15)	325.91 (%1)
	300	Average of AIC (% lowest AIC)	584.66 (%15.03)	586.16 (%13.02)	586.75 (%13.62)	584.58 (%15.23)	584.93 (%18.03)	584.31 (%14.82)	589.11 (%10.22)
		Average of BIC (% lowest BIC)	614.29 (%16.03)	615.79 (%14.82)	620.08 (%11.82)	614.21 (%16.83)	614.56 (%18.63)	617.65 (%13.02)	622.44 (%8.81)