Skip to main content
. 2021 Mar 10;159:107217. doi: 10.1016/j.csda.2021.107217

Table 7.

Percentage of number clusters identified correctly (IC) in loglinear models based on 1000 simulation replications when data are generated from (28).

τ c n0=50
n0=100
k=2
k=3
k=2
k=3
AIC BIC AIC BIC AIC BIC AIC BIC
0.5 10 1.6 91.7 0.3 90.9 1.6 94.9 0.4 93.9
20 0.1 71.2 0.0 68.7 0.0 80.1 0.0 77.1
1.0 10 2.1 92.5 1.2 92.1 1.3 95.3 0.6 95.2
20 0.0 78.1 0.0 76.5 0.1 83.4 0.0 83.0