Table 1.

PPL criterion for the three models fit to the growth hormone data: Selection model (SM), Mixture model 1 (MM1), and Mixture Model 2 (MM2) for four choices of T(r, r ∘ y). C_∞ is the criterion with k = ∞ and GOF is the goodness of fit component of the criterion.

Model	GOF	Complexity	C ∞
T(r, r∘y) = rJyJ − r1y1
SM	2960.2	2907.6	5867.8
MM1	2961.7	3958.6	6920.3
MM2	3058.3	3498.5	6556.8
T(r, r∘y) = rJ(rJyJ − r1y1)
SM	390.7	425.2	815.9
MM1	390.2	517.8	907.9
MM2	484.7	605.7	1090.3
$T(\mathit{r},\mathit{r}\circ \mathit{y})={\sum }_{j=1}^J\left\{\mathrm{I}(r_j=1,r_{j+1}=0)r_j{y}_j\right\}-\mathrm{I}(r_2=1)r_1{y}_1$
SM	1670.5	1759.7	3430.2
MM1	1670.0	2211.4	3881.4
MM2	1768.1	2606.3	4374.4
$T(\mathit{r},\mathit{r}\circ \mathit{y})={\left[{\sum }_{j=1}^J\left\{\mathrm{I}(r_j=1,r_{j+1}=0)r_j{y}_j\right\}-\mathrm{I}(r_2=1)r_1{y}_1\right]}^2$
SM	15563039	11655064	27218103
MM1	15712294	23472467	39184760
MM2	15760469	22043555	37804025