Modelling endogenous insulin concentration in type 2 diabetes during closed-loop insulin delivery

Table 2 Results of DIC analysis for the six competing models

Model	\( \overline{D} \)	\( \widehat{D} \)	p _D	*DIC*	d
1	9003	8982	21	9024	5597
2	6435	6409	26	6461	3034
3	4482	4435	47	4529	1102
4	4079	4024	55	4134	707
5	3394	3335	59	3453	26
6	3357	3287	70	3427	0

\( \overline{D}= \) posterior mean of -2log(likelihood); \( \widehat{D}= \) 2log(likelihood) at posterior mean of stochastic nodes; \( {p}_D=\kern0.5em \overline{D}-\widehat{D} \), defines “effective number of parameters”; \( DIC=\overline{D} + {p}_D \), is deviance information criteria. d is the difference between the present model’s DIC and the lowest DIC.

ISSN: 1475-925X