Update 5/6: Business case with test on i+1
Update 4/24: No tuning, all tunings, and top tunings
To complicate things a little more, let's add another variable!
- Curiously, I was unable to replicate the previous results without parameter tuning
- Using paramaterless Gaussian Bayes only, there is little difference between HI
- I repeated using parameter tuning, but calculating stats based on ALL results rather than only top-ranked results
- These are using 30 random train/test splits, but...
- These are only using 2 out of 3 learners to save param tune time
- 3-learner results can come later, but from what I've seen, 2vs3 doesn't matter
- Results with top-ranked parameters > results with all parameters > results with no parameters
Results with No Parameters:
Label , A12, U, p, meanA, meanB
HI 4 > HI 0 , 0.000, 0, 0.500000, 0.771908, 0.680063
HI 4 > HI 1 , 0.000, 0, 0.500000, 0.771908, 0.758529
HI 4 > HI 2 , 0.000, 0, 0.500000, 0.771908, 0.758749
HI 4 > HI 3 , 1.000, 0, 0.500000, 0.771908, 0.790568
HI 3 > HI 0 , 0.600, 11, 0.417266, 0.574222, 0.547370
HI 3 > HI 1 , 0.560, 9, 0.265435, 0.574222, 0.595885
HI 3 > HI 2 , 0.320, 9, 0.265435, 0.574222, 0.582948
HI 2 > HI 0 , 0.281, 85, 0.282867, 0.601895, 0.561865
HI 2 > HI 1 , 0.862, 90, 0.365195, 0.601895, 0.595265
HI 1 > HI 0 , 0.023, 216, 0.145822, 0.596752, 0.544316
Results with All Parameters:
Label , A12, U, p, meanA, meanB
HI 4 > HI 0 , 0.000, 0, 0.500000, 0.730214, 0.677388
HI 4 > HI 1 , 1.000, 0, 0.500000, 0.730214, 0.745145
HI 4 > HI 2 , 1.000, 0, 0.500000, 0.730214, 0.754882
HI 4 > HI 3 , 1.000, 0, 0.500000, 0.730214, 0.750597
HI 3 > HI 0 , 0.200, 8, 0.201698, 0.589507, 0.527802
HI 3 > HI 1 , 0.360, 11, 0.417266, 0.589507, 0.619794
HI 3 > HI 2 , 0.400, 12, 0.500000, 0.589507, 0.609074
HI 2 > HI 0 , 0.281, 74, 0.140122, 0.623904, 0.552588
HI 2 > HI 1 , 0.699, 97, 0.490836, 0.623904, 0.618135
HI 1 > HI 0 , 0.234, 188, 0.047494, 0.618359, 0.550335
Results with Top-Ranked Parameters:
HI 4 > HI 0 , 0.000, 0, 0.500000, 0.909430, 0.766430
HI 4 > HI 1 , 0.000, 0, 0.500000, 0.909430, 0.869284
HI 4 > HI 2 , 0.000, 0, 0.500000, 0.909430, 0.857104
HI 4 > HI 3 , 1.000, 0, 0.500000, 0.909430, 0.911470
HI 3 > HI 0 , 0.160, 6, 0.105038, 0.905671, 0.787974
HI 3 > HI 1 , 0.160, 5, 0.071836, 0.905671, 0.808840
HI 3 > HI 2 , 0.160, 7, 0.148135, 0.905671, 0.824976
HI 2 > HI 0 , 0.066, 64, 0.061872, 0.831493, 0.760456
HI 2 > HI 1 , 0.071, 85, 0.282867, 0.831493, 0.801677
HI 1 > HI 0 , 0.119, 192, 0.056850, 0.788835, 0.734083
Original Post
OK, to start off with, HI = History Index = number of past deltas included
- ant 1.7 with HI=3 would include deltas from ant 1.6, and 1.5, and ant 1.4
- ant 1.7 with HI=0 would included no deltas (just the original set)
The results below come from comparing only the top-ranked param tuning results on each delta
No comments:
Post a Comment