Table 1

AUC for different models on both simulated datasets.

TileMap

Baum-Welch

Viterbi-Training

Viterbi-EM

ad hoc

dataset I

0.9986

0.9998

0.9997

0.9998

0.9869

dataset II

0.9749

0.9995

0.9994

0.9995

0.9728

All models with optimised parameters outperform TileMap on both simulated datasets. While TileMap performs well on dataset I it is only slightly better than the model with ad hoc parameter estimates.

Humburg et al. BMC Bioinformatics 2008 9:343   doi:10.1186/1471-2105-9-343