February 14, 2005
More improvements to Strawman
These results reflect one change: I smoothed the CPTs between EM iterations to eliminate zero entries.
| Model | Time-Slice Accuracy | Edit Distance | |
| Primary Activity | Any Activity | ||
| "single-trained" straw man (prior) | 84.4% (std 6.6). | 90.1% (std 5.1) | 16.2 std(7.4) |
| "full-trained" straw man (state-transitions) | 71.43% (std 5.3) | 77.71% (std 5.7) | 80.3 (std 9.5) |
| pinned straw man (partial labels) | 51.89% (std 6.8) | 53.38% (std 5.9) | 120.2 (std 12.47) |
| labelled straw man | 86.48% (std 4.1) | 91.57% (std 3.4) | 14.1 (std 5.7) |
Comments
Post a comment