Yails has a self testing utility module. This module enables the user to test the system in any problem giving him the possibility of trying several parameter settings and outputing statistically significant results in order to enable him to make the correct decisions. Bellow an example run of this module is presented :
YAILS - test
==========================
YAILS Evaluation Assistant
==========================
DATA SET ?
lymp_stat.
Number of RUNS ? 10.
Percentage of Examples used for TRAINING (0..1) ? 0.7.
** Learning Parameters used on all runs **
Similarity Threshold (0..1) ? 0.5.
Minimal Consistency (0..1) ? 0.9.
Minimal Coverage (0..1) ? 1.
** Classification Tasks for each run **
-> Instructions :
0 - No more Tasks
1 - Normal Classification
2 - Foreground Theory Only
3 - All Rules Used (Fore+Back)
Type of Task (0,1,2 or 3) ? 1.
Minimal Utility (0..1) ? 0.05.
Minimal Matching (0..1) ? 0.5.
Type of Task (0,1,2 or 3) ? 3.
Minimal Matching (0..1) ? 0.5.
Type of Task (0,1,2 or 3) ? 0.
GENERATING TRAIN/TEST SPLITS
LEARNING THE TRAINING SETS
LEARNING THE TRAINING SET 10
LEARNING THE TRAINING SET 9
LEARNING THE TRAINING SET 8
....
....
CLASSIFYING THE TESTING SETS
normal(0.800,0.900)
all(0.900)
EVALUATING THE RESULTS
DATA SET
lymp_stat
Number of RUNS 10
Percentage of examples used for TRAINING 0.700
** Learning Parameters used on all runs **
Similarity Threshold 0.50
Minimal Consistency 0.90
Minimal Coverage 1.00
******************************
RESULTS AVERAGED OVER 10 RUNS
******************************
============================================================
TASK : all(0.500)
PARAMETER Mean Stand.Dev. Max. Min.
------------------------------------------------------------
%Accuracy 0.87 0.05 0.96 0.80
UsedRules 27.60 2.64 36.00 20.00
Avg.Sels/Rule 1.84 0.20 2.12 1.56
CorrectAnswers 39.00 2.45 43.00 36.00
IncorrectAnswers 6.00 2.45 9.00 2.00
UnknownAnswers 0.00 0.00 0.00 0.00
BackRules 14.70 2.21 19.00 12.00
============================================================
TASK : normal(0.050,0.500)
PARAMETER Mean Stand.Dev. Max. Min.
------------------------------------------------------------
%Accuracy 0.85 0.05 0.91 0.78
UsedRules 14.00 2.67 19.00 9.00
Avg.Sels/Rule 1.85 0.21 2.12 1.56
CorrectAnswers 38.20 2.15 41.00 35.00
IncorrectAnswers 6.80 2.15 10.00 4.00
UnknownAnswers 0.00 0.00 0.00 0.00
BackRules 13.60 2.50 19.00 10.00
As it can be seen this module enables the user to define the type of test he wants to make (the number of repetitions, the percentage of examples used for learning, the learning parameters, etc.). The system also enables to use several types of different classification strategies on each run. At the end Yails presents several results of the test together with statistical indicators.