DeLong pairwise AUC tests
==========================

RF vs LR

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -1.9871, p-value = 0.04691
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.0582257790 -0.0003998269
sample estimates:
AUC of roc1 AUC of roc2 
  0.7075777   0.7237810 
 

RF vs SVM

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 3.7586, p-value = 0.0001709
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.04406068 0.14008531
sample estimates:
AUC of roc1 AUC of roc2 
  0.7075777   0.6023952 
 

RF vs DT

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 4.2071, p-value = 2.587e-05
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.05131461 0.14082967
sample estimates:
AUC of roc1 AUC of roc2 
 0.7075777   0.5983961 
 

RF vs NB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 0.21541, p-value = 0.8295
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.03094537  0.03858723
sample estimates:
AUC of roc1 AUC of roc2 
 0.7075777   0.6906473 
 

RF vs KNN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 3.1661, p-value = 0.001545
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.03036682 0.12905719
sample estimates:
AUC of roc1 AUC of roc2 
  0.7075777   0.6147562 
 

RF vs XGB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 1.5593, p-value = 0.1189
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.005901857  0.051838555
sample estimates:
AUC of roc1 AUC of roc2 
  0.7075777   0.6714999 
 

RF vs ANN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 2.7074, p-value = 0.006781
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.01814885 0.11333105
sample estimates:
AUC of roc1 AUC of roc2 
  0.7075777   0.6287283 
 

RF vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -1.7632, p-value = 0.07787
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.054521294  0.002881716
sample estimates:
AUC of roc1 AUC of roc2 
  0.7075777   0.7202880 
 

RF vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 2.1405, p-value = 0.03232
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.003343996 0.075954550
sample estimates:
AUC of roc1 AUC of roc2 
  0.7075777   0.6548189 
 

LR vs SVM

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 4.45, p-value = 8.586e-06
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.06792275 0.17484885
sample estimates:
AUC of roc1 AUC of roc2 
  0.7237810   0.6023952 
 

LR vs DT

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 5.1412, p-value = 2.73e-07
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.07758451 0.17318538
sample estimates:
AUC of roc1 AUC of roc2 
  0.7237810   0.5983961 
 

LR vs NB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 1.8546, p-value = 0.06365
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.001881805  0.068149270
sample estimates:
AUC of roc1 AUC of roc2 
  0.7237810   0.6906473 
 

LR vs KNN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 4.0414, p-value = 5.314e-05
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.05615031 0.16189930
sample estimates:
AUC of roc1 AUC of roc2 
  0.7237810   0.6147562 
 

LR vs XGB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 2.5298, p-value = 0.01141
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.01177698 0.09278532
sample estimates:
AUC of roc1 AUC of roc2 
  0.7237810   0.6714999 
 

LR vs ANN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 3.8936, p-value = 9.878e-05
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.04720469 0.14290082
sample estimates:
AUC of roc1 AUC of roc2 
  0.7237810   0.6287283 
 

LR vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 0.77292, p-value = 0.4396
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.00536449  0.01235052
sample estimates:
AUC of roc1 AUC of roc2 
   0.723781    0.720288 
 

LR vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 3.9793, p-value = 6.913e-05
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.03499514 0.10292902
sample estimates:
AUC of roc1 AUC of roc2 
  0.7237810   0.6548189 
 

SVM vs DT

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 0.12251, p-value = 0.9025
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.05998254  0.06798083
sample estimates:
AUC of roc1 AUC of roc2 
  0.6023952   0.5983961 
 

SVM vs NB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -3.0987, p-value = 0.001944
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.14407297 -0.03243116
sample estimates:
AUC of roc1 AUC of roc2 
  0.6023952   0.6906473 
 

SVM vs KNN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -0.51561, p-value = 0.6061
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.05934811  0.03462613
sample estimates:
AUC of roc1 AUC of roc2 
  0.6023952   0.6147562 
 

SVM vs XGB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -2.476, p-value = 0.01329
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.12380721 -0.01440209
sample estimates:
AUC of roc1 AUC of roc2 
  0.6023952   0.6714999 
 

SVM vs ANN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -0.88435, p-value = 0.3765
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.08469438  0.03202829
sample estimates:
AUC of roc1 AUC of roc2 
  0.6023952   0.6287283 
 

SVM vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -4.3148, p-value = 1.597e-05
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.17144445 -0.06434112
sample estimates:
AUC of roc1 AUC of roc2 
  0.6023952   0.7202880 
 

SVM vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -1.856, p-value = 0.06345
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.107782612  0.002935164
sample estimates:
AUC of roc1 AUC of roc2 
  0.6023952   0.6548189 
 

DT vs NB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -3.4703, p-value = 0.0005199
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.14435309 -0.04014933
sample estimates:
AUC of roc1 AUC of roc2 
  0.5983961   0.6906473 
 

DT vs KNN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -0.53045, p-value = 0.5958
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.07680921  0.04408894
sample estimates:
AUC of roc1 AUC of roc2 
  0.5983961   0.6147562 
 

DT vs XGB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -2.8582, p-value = 0.00426
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.12323267 -0.02297491
sample estimates:
AUC of roc1 AUC of roc2 
  0.5983961   0.6714999 
 

DT vs ANN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -1.0428, p-value = 0.297
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.08734039  0.02667600
sample estimates:
AUC of roc1 AUC of roc2 
  0.5983961   0.6287283 
 

DT vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -4.9761, p-value = 6.488e-07
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.16990228 -0.07388158
sample estimates:
AUC of roc1 AUC of roc2 
  0.5983961   0.7202880 
 

DT vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -2.0245, p-value = 0.04292
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.111046535 -0.001799202
sample estimates:
AUC of roc1 AUC of roc2 
  0.5983961   0.6548189 
 

NB vs KNN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 2.6335, p-value = 0.008452
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.01940899 0.13237316
sample estimates:
AUC of roc1 AUC of roc2 
  0.6906473   0.6147562 
 

NB vs XGB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 0.86456, p-value = 0.3873
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.02425973  0.06255457
sample estimates:
AUC of roc1 AUC of roc2 
  0.6906473   0.6714999 
 

NB vs ANN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 2.3484, p-value = 0.01885
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.010242 0.113596
sample estimates:
AUC of roc1 AUC of roc2 
  0.6906473   0.6287283 
 

NB vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -1.6587, p-value = 0.09718
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.064665231  0.005383794
sample estimates:
AUC of roc1 AUC of roc2 
  0.6906473   0.7202880 
 

NB vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 1.6751, p-value = 0.09392
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.006093227  0.077749914
sample estimates:
AUC of roc1 AUC of roc2 
  0.6906473   0.6548189 
 

KNN vs XGB

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -2.0406, p-value = 0.0413
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.111246355 -0.002240957
sample estimates:
AUC of roc1 AUC of roc2 
  0.6147562   0.6714999 
 

KNN vs ANN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -0.46968, p-value = 0.6386
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.07227685  0.04433274
sample estimates:
AUC of roc1 AUC of roc2 
  0.6147562   0.6287283 
 

KNN vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -3.8695, p-value = 0.000109
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.15898471 -0.05207888
sample estimates:
AUC of roc1 AUC of roc2 
  0.6147562   0.7202880 
 

KNN vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -1.4681, p-value = 0.1421
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.09354823  0.01342277
sample estimates:
AUC of roc1 AUC of roc2 
  0.6147562   0.6548189 
 

XGB vs ANN

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 1.6053, p-value = 0.1084
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.009450518  0.094993717
sample estimates:
AUC of roc1 AUC of roc2 
  0.6714999   0.6287283 
 

XGB vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -2.3997, p-value = 0.01641
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.088635801 -0.008940475
sample estimates:
AUC of roc1 AUC of roc2 
  0.6714999   0.7202880 
 

XGB vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 0.75219, p-value = 0.4519
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.02678401  0.06014586
sample estimates:
AUC of roc1 AUC of roc2 
  0.6714999   0.6548189 
 

ANN vs LDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -3.7121, p-value = 0.0002056
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.13990305 -0.04321643
sample estimates:
AUC of roc1 AUC of roc2 
  0.6287283   0.7202880 
 

ANN vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = -1.0268, p-value = 0.3045
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 -0.07589505  0.02371370
sample estimates:
AUC of roc1 AUC of roc2 
  0.6287283   0.6548189 
 

LDA vs QDA

	DeLong's test for two correlated ROC curves

data:  model_rocs[[i]] and model_rocs[[j]]
Z = 3.8672, p-value = 0.0001101
alternative hypothesis: true difference in AUC is not equal to 0
95 percent confidence interval:
 0.03228788 0.09865024
sample estimates:
AUC of roc1 AUC of roc2 
  0.7202880   0.6548189 
 
