Table A.7.
POV | # of components | Naïve Bayes
|
SVM (linear kernel)
|
SVM (RBF kernel)
|
MaxEnt
|
||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
P. | R. | F1 | P. | R. | F1 | P. | R. | F1 | P. | R. | F1 | ||
| |||||||||||||
0.99 | 405 | 0.533 | 0.950 | 0.683 | 0.748 | 0.779 | 0.758 | 0.718 | 0.775 | 0.741 | 0.726 | 0.810 | 0.763 |
0.95 | 289 | 0.539 | 0.938 | 0.684 | 0.719 | 0.767 | 0.738 | 0.717 | 0.744 | 0.725 | 0.719 | 0.790 | 0.749 |
0.90 | 210 | 0.538 | 0.942 | 0.684 | 0.710 | 0.751 | 0.725 | 0.703 | 0.736 | 0.713 | 0.704 | 0.775 | 0.734 |
0.85 | 158 | 0.535 | 0.934 | 0.680 | 0.695 | 0.729 | 0.708 | 0.688 | 0.728 | 0.698 | 0.690 | 0.760 | 0.722 |
0.80 | 121 | 0.537 | 0.934 | 0.681 | 0.679 | 0.740 | 0.706 | 0.675 | 0.732 | 0.695 | 0.686 | 0.768 | 0.722 |
0.75 | 93 | 0.537 | 0.926 | 0.680 | 0.682 | 0.756 | 0.715 | 0.641 | 0.736 | 0.681 | 0.674 | 0.764 | 0.715 |
0.70 | 71 | 0.539 | 0.915 | 0.678 | 0.683 | 0.764 | 0.720 | 0.651 | 0.725 | 0.681 | 0.671 | 0.752 | 0.708 |
0.65 | 54 | 0.544 | 0.896 | 0.676 | 0.676 | 0.744 | 0.706 | 0.637 | 0.741 | 0.682 | 0.664 | 0.744 | 0.701 |
0.60 | 41 | 0.556 | 0.888 | 0.683 | 0.669 | 0.764 | 0.711 | 0.645 | 0.772 | 0.700 | 0.661 | 0.748 | 0.701 |
Note: Performance of shaded region is not significantly different from the top overall model (i.e., F1 of 77.4%).