5 Output files

During training phase, RuleKit produces following types of files:

• a model file (one per each training set) that can be applied in the prediction stage,
• a text report (common for all training files).

The result of the prediction phase are:

• a prediction file (one per each testing set),
• a performance report (common for all testing files). In the following subsections, a detailed description of training and performance reports are given.

5.1. Training report

The report consists of separated sections, each corresponding to a single traning file:

================================================================================
bmt-train-0.arff

... # content

================================================================================
bmt-train-1.arff

... # content

At the beginning of a section, induction parameters are given:

Params:
min_rule_covered=5.0
induction_measure=LogRankStatistics
pruning_measure=LogRankStatistics
voting_measure=LogRankStatistics

Next, a rule model is presented:

Rules:
r1: IF Relapse = {0} AND Donorage = (-inf, 45.526027) AND Recipientage = (-inf, 17.45) THEN survival_status = {NaN} (p=119.0, n=0.0, P=168.0, N=0.0, weight=0.9999992726837377, pval=7.27316262327804E-7)
r2: IF HLAmismatch = {0} AND Relapse = {1} THEN survival_status = {NaN} (p=21.0, n=0.0, P=168.0, N=0.0, weight=0.9981544870337137, pval=0.0018455129662863223)
r3: IF Relapse = {0} AND Rbodymass = (-inf, 69.0) AND Recipientage = (-inf, 18.0) THEN survival_status = {NaN} (p=127.0, n=0.0, P=168.0, N=0.0, weight=0.9999999653103507, pval=3.468964926423013E-8)
r4: IF aGvHDIIIIV = {1} AND ANCrecovery = (-inf, 19.5) AND Stemcellsource = {1} AND Txpostrelapse = {0} THEN survival_status = {NaN} (p=82.0, n=0.0, P=168.0, N=0.0, weight=0.999992179496458, pval=7.820503541977608E-6)
r5: IF Donorage = <28.028767000000002, inf) AND CD34kgx10d6 = <1.2650000000000001, 6.720000000000001) AND CD3dCD34 = <0.8878985, inf) AND Rbodymass = <31.5, inf) AND Recipientage = <11.55, inf) THEN survival_status = {NaN} (p=20.0, n=0.0, P=168.0, N=0.0, weight=0.9999999999914838, pval=8.516187754992188E-12)

For each rule, additional statistics are given in the parentheses:

• elements of confusion matrix p, n, P, N,
• weight - value of the voting quality measure,
• p-value - rule significance.

Rules are followed by the detailed information about training set coverage:

Coverage of training examples by rules (1-based):
2,4*;2*;3,5*;1,3*,4;1,3*;1,3*;1,3*;1,3*,4;1,3*;1,3*,4;3*,4;1,3,5*;1,3*;2*;1,3*,4;1,3*,4;2*;1,3*,4;1,3,5*;1,3*;1*;1,3*,4;3*,4;4,5*;3*,4;1,3*,4;1,3*,4;1,3*,4;1,3*,4;1,3*;1*,4;1,3*;1,3*;3*;1,3*,4;1,3*;1,3*,4;5*;1,3*,4;2*;3,5*;1,3*,4;1,3*;2,5*;1,3*,4;1,3*,4;3*,4;1,3*;1,3*;1,3*,4;2,4*;1,3*,4;4,5*;1,3*;2*;1,3*;2*;1,3*,4;1,3*;1,3*,4;3*;5*;1,3*,4;1,3*,4;1,3*;1,3*;1,3*,4;3,4,5*;4*;1,3*,4;1,3*;1,3*;1,3*,4;5*;4*;1,3*,4;1,3*,4;2*;1,3*;1,3*;1,3*;1,3*;1,3*;1,3*;1,3*,4;1,3*,4;1,3*;1,3*;1,3,5*;1,3*;1,3*,4;5*;1,3*;2*;1,3*;1*,4;1,3*,4;1,3*,4;1,3*,4;1,5*;1,3*;1,3*;4*;1,3*;3,5*;3*;1,3,4,5*;1,3*;5*;2*;1,3*;1,3*;1*,4;1,3*;1,3*,4;1,3*;1,3*,4;3*,4;1,3*;1,3*,4;4*;1,3*,4;1,3*;1,3*,4;4*;1,3*;1,3*;1,3*;1,3*,4;1*;1,3*,4;1,3*,4;2*;1,3*,4;1,3*,4;1,3,5*;2,4*;1,3*;1,3*,4;1,3*,4;1,3*;2,4*;1,3*,4;3*,4;1,3*,4;1,3*;1,3*;1,3*;1,3*,4;1,3*,4;1,3*,4;1,3*;1,3*;5*;2,4*;2*;1,3,5*;2,4*;3*;1,3*;2,4*;1,3*,4;1,3*,4;1,3*,4;1,3*,4;2*;4*;2,4*

For each example from the training set, a comma-separated list of rules covering that example is specified. Best rule (one with highest weight is marked with asterisk, lists corresponding to consecutive examples are separated with semicolon. The record 2,4*;2*; at the beginning indicates that the first training example was covered by rules r2 and r4 (of which r4 was the best), while the second training example was covered by r2 only.

Another section of the training report applies to survival problems only and contains tabular representation of survival curves. The first column represents time, then there are survival estimates of the entire training set and induced rules. This can be used for visualization of the algorithm results.

Estimator:
time, entire-set, r1, r2, r3, r4, r5, 
6.0, 0.9940476190476191, 0.9915966386554622,1.0,0.9921259842519685,1.0,1.0,
10.0, 0.988095238095238, 0.9915966386554622,1.0,0.984251968503937,1.0,1.0,
11.0, 0.9821428571428571, 0.9831932773109244,1.0,0.9763779527559056,1.0,1.0,
15.0, 0.976190476190476, 0.9831932773109244,1.0,0.9763779527559056,1.0,0.95,
19.0, 0.9702380952380951, 0.9831932773109244,1.0,0.9763779527559056,1.0,0.8999999999999999,
26.0, 0.9642857142857142, 0.9747899159663866,1.0,0.9685039370078741,1.0,0.8999999999999999,
28.0, 0.9523809523809522, 0.957983193277311,1.0,0.9527559055118111,0.9878048780487805,0.8999999999999999,
31.0, 0.9464285714285713, 0.9495798319327731,1.0,0.9448818897637796,0.9878048780487805,0.8999999999999999,
35.0, 0.9404761904761904, 0.9411764705882353,1.0,0.9370078740157481,0.9878048780487805,0.8999999999999999,
41.0, 0.9285714285714285, 0.9327731092436975,1.0,0.9291338582677167,0.975609756097561,0.8499999999999999,
42.0, 0.9226190476190476, 0.9243697478991597,1.0,0.9212598425196852,0.975609756097561,0.7999999999999998,
48.0, 0.9166666666666666, 0.9159663865546219,1.0,0.9212598425196852,0.975609756097561,0.7499999999999998,
53.0, 0.9107142857142857, 0.9159663865546219,1.0,0.9212598425196852,0.975609756097561,0.6999999999999998,
55.0, 0.9047619047619048, 0.9159663865546219,1.0,0.9133858267716537,0.975609756097561,0.6499999999999999,
56.0, 0.8988095238095238, 0.9159663865546219,1.0,0.9133858267716537,0.9634146341463414,0.6499999999999999,

The last element of the report are model indicators followed by the performance metrices evaluated on the training set. The contents of this section depends on the investigated problem.

Model characteristics:
time_total_s: 13.829900798
time_growing_s: 11.434164728999999
time_pruning_s: 2.3417725849999997
#rules: 5.0
#conditions_per_rule: 3.6
#induced_conditions_per_rule: 73.6
avg_rule_coverage: 0.43928571428571433
avg_rule_precision: 1.0
avg_rule_quality: 0.9996291809031488
avg_pvalue: 3.7081909685121595E-4
avg_FDR_pvalue: 3.71317511237688E-4
avg_FWER_pvalue: 3.726949446670513E-4
fraction_0.05_significant: 1.0
fraction_0.05_FDR_significant: 1.0
fraction_0.05_FWER_significant: 1.0

Training set performance:
integrated_brier_score: 0.20504955695116336

5.2. Prediction performance report

The prediction performance report has the form of comma-separated table with rows corresponding to testing sets and columns representing model indicators and, optionally, performance metrices. The latter are reported only when real labels are provided in the testing set.

Parameters: min_rule_covered=5.0; induction_measure=LogRankStatistics; pruning_measure=LogRankStatistics; voting_measure=LogRankStatistics; 
Dataset, time started, elapsed[s], time_total_s,time_growing_s,time_pruning_s,#rules,#conditions_per_rule,#induced_conditions_per_rule,avg_rule_coverage,avg_rule_precision,avg_rule_quality,avg_pvalue,avg_FDR_pvalue,avg_FWER_pvalue,fraction_0.05_significant,fraction_0.05_FDR_significant,fraction_0.05_FWER_significant,integrated_brier_score,
bmt-test-0.arff,2018.10.10_19.22.19,14.197630903,13.829900798, 11.434164728999999, 2.3417725849999997, 5.0, 3.6, 73.6, 0.43928571428571433, 1.0, 0.9996291809031488, 3.7081909685121595E-4, 3.71317511237688E-4, 3.726949446670513E-4, 1.0, 1.0, 1.0, 0.20866685101468796, 
bmt-test-1.arff,2018.10.10_19.22.33,8.816920328,8.785011626, 7.139117646, 1.644193643, 3.0, 2.0, 81.33333333333333, 0.5277777777777778, 1.0, 0.9999840894880433, 1.591051195670712E-5, 1.593935102948511E-5, 1.5968190102263097E-5, 1.0, 1.0, 1.0, 0.32738956730627355, 
bmt-test-2.arff,2018.10.10_19.22.42,11.202350452,11.139475589, 9.226062807, 1.9120130830000002, 5.0, 3.4, 65.0, 0.4238095238095238, 1.0, 0.995274964099923, 0.004725035900076935, 0.004725037058680282, 0.004725040534490322, 1.0, 1.0, 1.0, 0.2193283512681922, 
bmt-test-3.arff,2018.10.10_19.22.53,11.235637043,11.117153526, 9.119976444, 1.995392923, 7.0, 3.857142857142857, 60.57142857142857, 0.35289115646258506, 1.0, 0.9801963244964088, 0.019803675503591172, 0.01986258037321523, 0.020129896761985813, 0.8571428571428571, 0.8571428571428571, 0.8571428571428571, 0.1969380949017121, 
bmt-test-4.arff,2018.10.10_19.23.04,10.668216899,10.551245553, 8.727900207000001, 1.822214638, 4.0, 2.25, 67.75, 0.49999999999999994, 1.0, 0.9999999359313667, 6.406863328756174E-8, 6.406863328756174E-8, 6.406863328756174E-8, 1.0, 1.0, 1.0, 0.17517417672611418, 
bmt-test-5.arff,2018.10.10_19.23.15,10.523336237,10.4100414, 8.722353154, 1.6866635950000002, 4.0, 4.0, 76.0, 0.4821428571428571, 1.0, 0.9781969933568722, 0.021803006643127898, 0.02181038208090318, 0.021823023749909604, 0.75, 0.75, 0.75, 0.25878170984273013, 
bmt-test-6.arff,2018.10.10_19.23.25,12.877260506,12.7574517, 10.616004643, 2.140288125, 5.0, 3.2, 77.6, 0.39880952380952384, 1.0, 0.9995242324316976, 4.7576756830243206E-4, 4.805232398770713E-4, 4.946063707460313E-4, 1.0, 1.0, 1.0, 0.19484740108158835, 
bmt-test-7.arff,2018.10.10_19.23.38,17.133387559,17.01621479, 14.541448512999997, 2.473219469, 7.0, 4.571428571428571, 63.57142857142857, 0.23245984784446325, 1.0, 0.9961646528405262, 0.003835347159473836, 0.004022999396535294, 0.004865332297101272, 1.0, 1.0, 1.0, 0.1759002883753468, 
bmt-test-8.arff,2018.10.10_19.23.55,22.294961552,22.183485085, 18.778375772, 3.403339693, 9.0, 4.777777777777778, 67.11111111111111, 0.23537146614069693, 1.0, 0.9949077503124099, 0.005092249687590009, 0.005743702908769271, 0.009426945617211135, 1.0, 1.0, 1.0, 0.16894150945629255, 
bmt-test-9.arff,2018.10.10_19.24.18,7.400985905,7.287232628, 6.234206981, 1.052146398, 3.0, 3.0, 61.333333333333336, 0.35897435897435903, 1.0, 0.9991030527299692, 8.969472700306828E-4, 8.969472700306828E-4, 8.969472700306828E-4, 1.0, 1.0, 1.0, 0.19829001480313704,