Table 2 Improvements of cBW + model selection, cBW alon vs Scheffer et al., with fully connected initial transition matrix for synthetic data with Viterbi algorithm
From: A new algorithm to train hidden Markov models for biological sequences with partial labels
State #/training sample | Average improvements of cBW + model selection in setting 1/2 | Average p value of cBW + model selection in setting 1/2 | Average improvements of cBW alone in setting 1/2 |
|---|---|---|---|
3/1600 | 7.35/8.29% | 2.1E−02/6.2E−05 | 7.61/8.49% |
3/2000 | 7.99/8.82% | 3.9E−02/2.6E−09 | 8.31/9.00% |
3/2400 | 8.47/9.13% | 2.8E−03/2.4E−10 | 8.71/9.18% |
3/2800 | 8.51/9.24% | 5.8E−03/6.9E−10 | 8.65/9.24% |
5/1600 | 12.97/14.75% | 7.8E−05/3.3E−05 | 11.13/12.82% |
5/2000 | 14.63/16.32% | 2.5E−03/1.2E−06 | 12.65/14.11% |
5/2400 | 14.69/16.54% | 7.8E−04/6.2E−06 | 12.73/14.50% |
5/2800 | 15.56/17.22% | 1.6E−02/1.3E−07 | 13.72/15.22% |
7/1600 | 8.61/10.42% | 3.5E−02/1.1E−02 | 5.56/7.20% |
7/2000 | 10.56/12.40% | 6.4E−03/6.2E−03 | 7.87/9.52% |
7/2400 | 11.35/13.21% | 8.2E−03/7.8E−03 | 8.71/10.43% |
7/2800 | 12.16/14.06% | 1.0E−03/1.1E−04 | 9.68/11.41% |