FastTree: Instantiate feature map for disk transpose and make Generalized Additive Models predictor resilient when feature map is not available. (#122)

* Instantiate feature map for disk transpose and make Generalized Additive Models predictor resilient when feature map is not available.

Author: codemzs

ZBaselines/SingleDebug/BinaryClassificationGamTrainer/BinaryClassificationGamTrainer-TrainTest-breast-cancer-out.txt

@@ -0,0 +1,53 @@
+maml.exe TrainTest test=%Data% tr=BinaryClassificationGamTrainer dout=%Output% data=%Data% out=%Output% seed=1
+Not adding a normalizer.
+Making per-feature arrays
+Changing data from row-wise to column-wise
+Warning: Skipped 16 instances with missing features during training
+Processed 683 instances
+Binning and forming Feature objects
+Starting to train ...
+Training calibrator.
+TEST POSITIVE RATIO:	0.3448 (241.0/(241.0+458.0))
+Confusion table
+          ||======================
+PREDICTED || positive | negative | Recall
+TRUTH     ||======================
+ positive ||      227 |       14 | 0.9419
+ negative ||       13 |      445 | 0.9716
+          ||======================
+Precision ||   0.9458 |   0.9695 |
+OVERALL 0/1 ACCURACY: 0.961373
+LOG LOSS/instance:  0.145652
+Test-set entropy (prior Log-Loss/instance): 0.929318
+LOG-LOSS REDUCTION (RIG): 84.326961
+AUC:                0.991198
+
+OVERALL RESULTS
+---------------------------------------
+AUC:                0.991198 (0.0000)
+Accuracy:           0.961373 (0.0000)
+Positive precision: 0.945833 (0.0000)
+Positive recall:    0.941909 (0.0000)
+Negative precision: 0.969499 (0.0000)
+Negative recall:    0.971616 (0.0000)
+Log-loss:           0.145652 (0.0000)
+Log-loss reduction: 84.326961 (0.0000)
+F1 Score:           0.943867 (0.0000)
+AUPRC:              0.971819 (0.0000)
+
+---------------------------------------
+Physical memory usage(MB): %Number%
+Virtual memory usage(MB): %Number%
+%DateTime%	 Time elapsed(s): %Number%
+
+--- Progress log ---
+[1] 'FastTree data preparation' started.
+[1] 'FastTree data preparation' finished in %Time%.
+[2] 'FastTree in-memory bins initialization' started.
+[2] 'FastTree in-memory bins initialization' finished in %Time%.
+[3] 'FastTree feature conversion' started.
+[3] 'FastTree feature conversion' finished in %Time%.
+[4] 'GAM training' started.
+[4] 'GAM training' finished in %Time%.
+[5] 'Saving model' started.
+[5] 'Saving model' finished in %Time%.

ZBaselines/SingleDebug/BinaryClassificationGamTrainer/BinaryClassificationGamTrainer-TrainTest-breast-cancer-rp.txt

@@ -0,0 +1,4 @@
+BinaryClassificationGamTrainer
+AUC	Accuracy	Positive precision	Positive recall	Negative precision	Negative recall	Log-loss	Log-loss reduction	F1 Score	AUPRC	Learner Name	Train Dataset	Test Dataset	Results File	Run Time	Physical Memory	Virtual Memory	Command Line	Settings	
+0.991198	0.961373	0.945833	0.941909	0.969499	0.971616	0.145652	84.32696	0.943867	0.971819	BinaryClassificationGamTrainer	%Data%	%Data%	%Output%	99	0	0	maml.exe TrainTest test=%Data% tr=BinaryClassificationGamTrainer dout=%Output% data=%Data% out=%Output% seed=1		
+

ZBaselines/SingleDebug/BinaryClassificationGamTrainer/BinaryClassificationGamTrainer-TrainTest-breast-cancer.txt

@@ -0,0 +1,49 @@
+maml.exe TrainTest test=%Data% tr=BinaryClassificationGamTrainer{dt+} dout=%Output% data=%Data% out=%Output% seed=1
+Not adding a normalizer.
+Making per-feature arrays
+Changing data from row-wise to column-wise on disk
+Warning: 16 of 699 examples will be skipped due to missing feature values
+Processed 683 instances
+Binning and forming Feature objects
+Starting to train ...
+Training calibrator.
+TEST POSITIVE RATIO:	0.3448 (241.0/(241.0+458.0))
+Confusion table
+          ||======================
+PREDICTED || positive | negative | Recall
+TRUTH     ||======================
+ positive ||      227 |       14 | 0.9419
+ negative ||       13 |      445 | 0.9716
+          ||======================
+Precision ||   0.9458 |   0.9695 |
+OVERALL 0/1 ACCURACY: 0.961373
+LOG LOSS/instance:  0.145652
+Test-set entropy (prior Log-Loss/instance): 0.929318
+LOG-LOSS REDUCTION (RIG): 84.326961
+AUC:                0.991198
+
+OVERALL RESULTS
+---------------------------------------
+AUC:                0.991198 (0.0000)
+Accuracy:           0.961373 (0.0000)
+Positive precision: 0.945833 (0.0000)
+Positive recall:    0.941909 (0.0000)
+Negative precision: 0.969499 (0.0000)
+Negative recall:    0.971616 (0.0000)
+Log-loss:           0.145652 (0.0000)
+Log-loss reduction: 84.326961 (0.0000)
+F1 Score:           0.943867 (0.0000)
+AUPRC:              0.971819 (0.0000)
+
+---------------------------------------
+Physical memory usage(MB): %Number%
+Virtual memory usage(MB): %Number%
+%DateTime%	 Time elapsed(s): %Number%
+
+--- Progress log ---
+[1] 'FastTree disk-based bins initialization' started.
+[1] 'FastTree disk-based bins initialization' finished in %Time%.
+[2] 'GAM training' started.
+[2] 'GAM training' finished in %Time%.
+[3] 'Saving model' started.
+[3] 'Saving model' finished in %Time%.

ZBaselines/SingleDebug/BinaryClassificationGamTrainer/BinaryClassificationGamTrainerDiskTranspose-TrainTest-breast-cancer-out.txt

@@ -0,0 +1,4 @@
+BinaryClassificationGamTrainer
+AUC	Accuracy	Positive precision	Positive recall	Negative precision	Negative recall	Log-loss	Log-loss reduction	F1 Score	AUPRC	/dt	Learner Name	Train Dataset	Test Dataset	Results File	Run Time	Physical Memory	Virtual Memory	Command Line	Settings	
+0.991198	0.961373	0.945833	0.941909	0.969499	0.971616	0.145652	84.32696	0.943867	0.971819	+	BinaryClassificationGamTrainer	%Data%	%Data%	%Output%	99	0	0	maml.exe TrainTest test=%Data% tr=BinaryClassificationGamTrainer{dt+} dout=%Output% data=%Data% out=%Output% seed=1	/dt:+	
+