“Discriminative power”版本间的差异
来自cslt Wiki
(相同用户的12个中间修订版本未显示) | |||
第1行: | 第1行: | ||
− | disc: test data = train data, cut 0-1 | + | == disc: test data = train data, cut 0-1 == |
+ | |||
第29行: | 第30行: | ||
!max T||pattern||cost||test on 1-2m | !max T||pattern||cost||test on 1-2m | ||
|- | |- | ||
− | |2||[5,35]||15.7213|| | + | |2||[5,35]||15.7213|| 16.569790 |
+ | |- | ||
+ | |3 || [2 7 31] || 15.6671|| 16.883425 | ||
+ | |- | ||
+ | |4|| [2 5 17 16]|| 15.4019||16.555786 | ||
+ | |- | ||
+ | |5||[3 2 17 8 10] || 15.3566 ||16.770004 | ||
+ | |- | ||
+ | |6|| [0 2 4 2 0 32]|| 15.2871||17.374308 | ||
+ | |- | ||
+ | |7|| [0 4 5 4 3 22 2] || 15.1647||16.984064 | ||
+ | |- | ||
+ | |8|| [1 1 7 1 1 25 3 1] || 15.2093 || 17.207624 | ||
+ | |} | ||
+ | |||
+ | == heterogeneous: train 0-1, criteria 1-2 == | ||
+ | {| | ||
+ | !max T||pattern||cost | ||
+ | |- | ||
+ | |2||[5 35]||16.5698 | ||
|- | |- | ||
− | |3 ||[ | + | |3||[8 21 11]||16.5417 |
|- | |- | ||
− | |4||[ | + | |4||[3 19 11 7]||16.3637 |
|- | |- | ||
− | |5||[1 | + | |5||[7 21 1 1 10]||16.4906 |
|- | |- | ||
− | |6 | + | |6|| |
|- | |- | ||
− | |7||[ | + | |7||[3 17 4 6 6 2 2]||16.5634 |
|- | |- | ||
− | |||
|} | |} |
2012年9月29日 (六) 08:14的最后版本
disc: test data = train data, cut 0-1
homogeneous
T | cost | test on 1-2m |
---|---|---|
1 | 16.089787 | 16.880100 |
2 | 15.845197 | 16.754873 |
3 | 15.714764 | 16.895517 |
4 | 15.488785 | 17.119724 |
5 | 15.601522 | 17.504296 |
6 | 15.658465 | 17.794173 |
7 | 15.972915 | 18.307699 |
8 | 16.222829 | 18.544731 |
heterogeneous
max T | pattern | cost | test on 1-2m |
---|---|---|---|
2 | [5,35] | 15.7213 | 16.569790 |
3 | [2 7 31] | 15.6671 | 16.883425 |
4 | [2 5 17 16] | 15.4019 | 16.555786 |
5 | [3 2 17 8 10] | 15.3566 | 16.770004 |
6 | [0 2 4 2 0 32] | 15.2871 | 17.374308 |
7 | [0 4 5 4 3 22 2] | 15.1647 | 16.984064 |
8 | [1 1 7 1 1 25 3 1] | 15.2093 | 17.207624 |
heterogeneous: train 0-1, criteria 1-2
max T | pattern | cost |
---|---|---|
2 | [5 35] | 16.5698 |
3 | [8 21 11] | 16.5417 |
4 | [3 19 11 7] | 16.3637 |
5 | [7 21 1 1 10] | 16.4906 |
6 | ||
7 | [3 17 4 6 6 2 2] | 16.5634 |