Skip to main content
. Author manuscript; available in PMC: 2010 Oct 26.
Published in final edited form as: J Am Stat Assoc. 2007 Jun 1;102(478):560–572. doi: 10.1198/016214507000000112

Table 1.

Model Performance in Simulation Examples

Simulated data



Model Simulation scheme FA LA HA–Bin Mix FA LA HA–Bin Mix FA LA HA–Bin Mix
1(a) FA 50.1 57.3 51.4 53.6 55.1 68.2 57.8 57.1 105.2 125.5 109.2 110.7
LA 23.1 17.9 18.1 18.9 37.8 20.7 25.8 30.6 60.9 38.6 43.9 49.5
HA–Bin 44.2 43.7 29.5 31.5 39.9 38.7 31.6 35.1 84.1 82.4 61.1 66.6
Mix 62.7 58.6 53.1 50.2 69.4 68.7 63.8 64.9 132.1 127.3 116.9 115.1
1(c) FA 51.1 64.1 52.2 51.9 53.9 59.8 55.8 57.4 105.0 123.9 108.0 109.3
LA 27.7 19.5 19.8 19.7 27.5 21.6 24.0 24.4 55.2 41.1 43.8 44.1
HA–Bin 45.5 43.9 35.8 35.9 48.6 40.0 36.5 39.3 94.1 83.9 72.3 75.2
Mix 49.8 49.5 43.7 43.4 62.3 60.8 59.3 60.1 112.1 110.3 103.0 103.5
1(d) FA 71.6 72.7 72.0 71.2 68.7 83.1 70.7 78.1 140.3 155.8 142.7 149.3
LA 26.7 18.5 20.3 20.5 35.5 22.7 26.5 26.1 62.2 41.2 46.8 46.6
HA–Bin 43.7 47.8 29.6 33.5 43.5 43.3 31.3 35.6 87.2 91.1 60.9 68.1
Mix 67.3 69.8 60.6 61.9 86.9 91.7 82.3 80.6 154.2 161.5 142.9 142.5
2 FA 45.9 53.4 48.0 48.6 55.9 67.4 58.8 61.1 101.8 120.8 106.8 109.7
LA 31.6 25.1 26.1 25.2 34.4 29.9 35.1 34.5 66.0 55.0 61.2 59.7
HA–Bin 59.7 62.6 47.6 49.5 69.4 69.6 60.6 63.5 129.1 132.2 108.2 113.0
Mix 57.8 56.2 46.0 46.2 67.5 65.5 63.4 60.5 125.3 121.7 109.4 106.7

NOTE: The first and second columns denote the underlying model and mechanism that generated the data. The remaining columns show means of goodness of fit (), the penalty (), and the L measure () under first-, last-, and hierarchical-activation schemes. The cells with the lowest means for each measure are highlighted.