further cleanup, detailled pa-predictions separated, text modified until results

1 files changed, 364 insertions, 0 deletions

diff --git a/data.yaml b/data.yaml
new file mode 100644
index 0000000..a7bbe6e
--- /dev/null
+++ b/data.yaml
@@ -0,0 +1,364 @@
+---
+:cv:
+  lazar-mp2d-all:
+    :tp: 3334
+    :fp: 830
+    :tn: 3054
+    :fn: 572
+    :n: 7790
+    :acc: 0.82
+    :tpr: 0.85
+    :fpr: 0.21
+    :tnr: 0.79
+    :ppv: 0.8
+    :npv: 0.84
+    :acc_perc: 82
+    :tpr_perc: 85
+    :tnr_perc: 79
+    :ppv_perc: 80
+    :npv_perc: 84
+  lazar-mp2d-high-confidence:
+    :tp: 2807
+    :fp: 560
+    :tn: 2089
+    :fn: 352
+    :n: 5808
+    :acc: 0.84
+    :tpr: 0.89
+    :fpr: 0.21
+    :tnr: 0.79
+    :ppv: 0.83
+    :npv: 0.86
+    :acc_perc: 84
+    :tpr_perc: 89
+    :tnr_perc: 79
+    :ppv_perc: 83
+    :npv_perc: 86
+  lazar-cdk-all:
+    :tp: 700
+    :fp: 1583
+    :tn: 2410
+    :fn: 3390
+    :n: 8083
+    :acc: 0.38
+    :tpr: 0.17
+    :fpr: 0.4
+    :tnr: 0.6
+    :ppv: 0.31
+    :npv: 0.42
+    :acc_perc: 38
+    :tpr_perc: 17
+    :tnr_perc: 60
+    :ppv_perc: 31
+    :npv_perc: 42
+  lazar-cdk-high-confidence:
+    :tp: 700
+    :fp: 1583
+    :tn: 2410
+    :fn: 3390
+    :n: 8083
+    :acc: 0.38
+    :tpr: 0.17
+    :fpr: 0.4
+    :tnr: 0.6
+    :ppv: 0.31
+    :npv: 0.42
+    :acc_perc: 38
+    :tpr_perc: 17
+    :tnr_perc: 60
+    :ppv_perc: 31
+    :npv_perc: 42
+  tensorflow-rf-mp2d:
+    :tp: 3197
+    :fp: 893
+    :tn: 3458
+    :fn: 742
+    :n: 8290
+    :acc: 0.8
+    :tpr: 0.81
+    :fpr: 0.21
+    :tnr: 0.79
+    :ppv: 0.78
+    :npv: 0.82
+    :acc_perc: 80
+    :tpr_perc: 81
+    :tnr_perc: 79
+    :ppv_perc: 78
+    :npv_perc: 82
+  tensorflow-lr-mp2d:
+    :tp: 3414
+    :fp: 676
+    :tn: 3541
+    :fn: 659
+    :n: 8290
+    :acc: 0.84
+    :tpr: 0.84
+    :fpr: 0.16
+    :tnr: 0.84
+    :ppv: 0.83
+    :npv: 0.84
+    :acc_perc: 84
+    :tpr_perc: 84
+    :tnr_perc: 84
+    :ppv_perc: 83
+    :npv_perc: 84
+  tensorflow-lr2-mp2d:
+    :tp: 3398
+    :fp: 692
+    :tn: 3560
+    :fn: 640
+    :n: 8290
+    :acc: 0.84
+    :tpr: 0.84
+    :fpr: 0.16
+    :tnr: 0.84
+    :ppv: 0.83
+    :npv: 0.85
+    :acc_perc: 84
+    :tpr_perc: 84
+    :tnr_perc: 84
+    :ppv_perc: 83
+    :npv_perc: 85
+  tensorflow-nn-mp2d:
+    :tp: 3375
+    :fp: 715
+    :tn: 3585
+    :fn: 615
+    :n: 8290
+    :acc: 0.84
+    :tpr: 0.85
+    :fpr: 0.17
+    :tnr: 0.83
+    :ppv: 0.83
+    :npv: 0.85
+    :acc_perc: 84
+    :tpr_perc: 85
+    :tnr_perc: 83
+    :ppv_perc: 83
+    :npv_perc: 85
+  tensorflow-svm-mp2d:
+    :tp: 3390
+    :fp: 700
+    :tn: 3611
+    :fn: 589
+    :n: 8290
+    :acc: 0.84
+    :tpr: 0.85
+    :fpr: 0.16
+    :tnr: 0.84
+    :ppv: 0.83
+    :npv: 0.86
+    :acc_perc: 84
+    :tpr_perc: 85
+    :tnr_perc: 84
+    :ppv_perc: 83
+    :npv_perc: 86
+  tensorflow-rf-cdk:
+    :tp: 3273
+    :fp: 747
+    :tn: 3483
+    :fn: 562
+    :n: 8065
+    :acc: 0.84
+    :tpr: 0.85
+    :fpr: 0.18
+    :tnr: 0.82
+    :ppv: 0.81
+    :npv: 0.86
+    :acc_perc: 84
+    :tpr_perc: 85
+    :tnr_perc: 82
+    :ppv_perc: 81
+    :npv_perc: 86
+  tensorflow-lr-cdk:
+    :tp: 3252
+    :fp: 768
+    :tn: 3165
+    :fn: 880
+    :n: 8065
+    :acc: 0.8
+    :tpr: 0.79
+    :fpr: 0.2
+    :tnr: 0.8
+    :ppv: 0.81
+    :npv: 0.78
+    :acc_perc: 80
+    :tpr_perc: 79
+    :tnr_perc: 80
+    :ppv_perc: 81
+    :npv_perc: 78
+  tensorflow-lr2-cdk:
+    :tp: 3215
+    :fp: 805
+    :tn: 3251
+    :fn: 794
+    :n: 8065
+    :acc: 0.8
+    :tpr: 0.8
+    :fpr: 0.2
+    :tnr: 0.8
+    :ppv: 0.8
+    :npv: 0.8
+    :acc_perc: 80
+    :tpr_perc: 80
+    :tnr_perc: 80
+    :ppv_perc: 80
+    :npv_perc: 80
+  tensorflow-nn-cdk:
+    :tp: 3419
+    :fp: 601
+    :tn: 3435
+    :fn: 610
+    :n: 8065
+    :acc: 0.85
+    :tpr: 0.85
+    :fpr: 0.15
+    :tnr: 0.85
+    :ppv: 0.85
+    :npv: 0.85
+    :acc_perc: 85
+    :tpr_perc: 85
+    :tnr_perc: 85
+    :ppv_perc: 85
+    :npv_perc: 85
+  tensorflow-svm-cdk:
+    :tp: 3292
+    :fp: 728
+    :tn: 3332
+    :fn: 713
+    :n: 8065
+    :acc: 0.82
+    :tpr: 0.82
+    :fpr: 0.18
+    :tnr: 0.82
+    :ppv: 0.82
+    :npv: 0.82
+    :acc_perc: 82
+    :tpr_perc: 82
+    :tnr_perc: 82
+    :ppv_perc: 82
+    :npv_perc: 82
+  :n: 8309
+  :n_uniq: 8290
+  :cdk:
+    :n_descriptors: 1442
+    :n_compounds: 8083
+:pa:
+  :n: 602
+  :lazar_all_mp2d:
+    :n: 560
+    :mut: 111
+    :non_mut: 449
+    :n_perc: 93
+    :mut_perc: 20
+    :non_mut_perc: 80
+  :lazar_high_confidence_mp2d:
+    :n: 301
+    :mut: 76
+    :non_mut: 225
+    :n_perc: 50
+    :mut_perc: 25
+    :non_mut_perc: 75
+  :lazar_all_cdk:
+    :n: 560
+    :mut: 111
+    :non_mut: 449
+    :n_perc: 93
+    :mut_perc: 20
+    :non_mut_perc: 80
+  :lazar_high_confidence_cdk:
+    :n: 301
+    :mut: 76
+    :non_mut: 225
+    :n_perc: 50
+    :mut_perc: 25
+    :non_mut_perc: 75
+  :lr_sgd_mp2d:
+    :n: 602
+    :mut: 127
+    :non_mut: 475
+    :n_perc: 100
+    :mut_perc: 21
+    :non_mut_perc: 79
+  :lr_scikit_mp2d:
+    :n: 602
+    :mut: 118
+    :non_mut: 484
+    :n_perc: 100
+    :mut_perc: 20
+    :non_mut_perc: 80
+  :nn_mp2d:
+    :n: 602
+    :mut: 124
+    :non_mut: 478
+    :n_perc: 100
+    :mut_perc: 21
+    :non_mut_perc: 79
+  :rf_mp2d:
+    :n: 602
+    :mut: 28
+    :non_mut: 574
+    :n_perc: 100
+    :mut_perc: 5
+    :non_mut_perc: 95
+  :svm_mp2d:
+    :n: 602
+    :mut: 78
+    :non_mut: 524
+    :n_perc: 100
+    :mut_perc: 13
+    :non_mut_perc: 87
+  :lr_sgd_cdk:
+    :n: 602
+    :mut: 97
+    :non_mut: 505
+    :n_perc: 100
+    :mut_perc: 16
+    :non_mut_perc: 84
+  :lr_scikit_cdk:
+    :n: 602
+    :mut: 88
+    :non_mut: 514
+    :n_perc: 100
+    :mut_perc: 15
+    :non_mut_perc: 85
+  :nn_cdk:
+    :n: 602
+    :mut: 150
+    :non_mut: 452
+    :n_perc: 100
+    :mut_perc: 25
+    :non_mut_perc: 75
+  :rf_cdk:
+    :n: 602
+    :mut: 10
+    :non_mut: 592
+    :n_perc: 100
+    :mut_perc: 2
+    :non_mut_perc: 98
+  :svm_cdk:
+    :n: 602
+    :mut: 18
+    :non_mut: 584
+    :n_perc: 100
+    :mut_perc: 3
+    :non_mut_perc: 97
+  :groups:
+    :Retronecine:
+      :n: 392
+    :Otonecine:
+      :n: 46
+    :Platynecine:
+      :n: 140
+    :N_oxide:
+      :n: 84
+    :Dehydropyrrolizidine:
+      :n: 23
+    :Tertiary_PA:
+      :n: 495
+    :Macrocyclic_diester:
+      :n: 255
+    :Monoester:
+      :n: 154
+    :Diester:
+      :n: 163