Species 123: Pinus pungens

General Linear Models

Boosted Regression Trees

Generalized Linear Model Results

Data:

C:\Withrow\Workspace\Model\glm_1\CovariateCorrelationOutputMDS_initial.csv

n(pres)                 =      233

n(abs)                  =      4427

n covariates considered =      5

total time for model fitting = 0.3min

Settings:

model family          : binomial

simplification method : AIC

Results:

number covariates in final model   : 13

Call:

glm(formula = response ~ bio16 + I(di_all_model^2) + I(bio16^2) + 

    di_all_model + I(bio15^2) + I(bio01^2) + bio02 + bio15 + 

    bio01 + bio16:bio15 + bio16:bio01 + bio16:bio02 + di_all_model:bio01, 

    family = model.family, data = dat, weights = weight, na.action = "na.exclude")

Deviance Residuals: 

    Min       1Q   Median       3Q      Max  

-2.4309  -0.1896  -0.0774  -0.0149   3.4566  

Coefficients:

                     Estimate Std. Error z value Pr(>|z|)    

(Intercept)         1.613e+01  2.059e+01   0.784 0.433268    

bio16              -1.701e-01  6.758e-02  -2.517 0.011843 *  

I(di_all_model^2)  -7.080e-03  1.443e-03  -4.905 9.34e-07 ***

I(bio16^2)          2.568e-04  3.603e-05   7.127 1.03e-12 ***

di_all_model       -1.112e-01  1.471e-01  -0.756 0.449775    

I(bio15^2)         -4.670e-02  1.871e-02  -2.496 0.012576 *  

I(bio01^2)         -1.168e-03  4.066e-04  -2.873 0.004066 ** 

bio02               7.559e-01  1.499e-01   5.041 4.63e-07 ***

bio15              -5.110e-01  7.324e-01  -0.698 0.485369    

bio01              -4.511e-01  1.170e-01  -3.856 0.000115 ***

bio16:bio15         3.961e-03  1.327e-03   2.985 0.002834 ** 

bio16:bio01         1.293e-03  2.266e-04   5.706 1.16e-08 ***

bio16:bio02        -1.851e-03  4.818e-04  -3.843 0.000122 ***

di_all_model:bio01  4.563e-03  1.629e-03   2.801 0.005094 ** 

---

Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 1850.16  on 4659  degrees of freedom

Residual deviance:  974.79  on 4646  degrees of freedom

AIC: 1002.8

Number of Fisher Scoring iterations: 9

============================================================ 

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6180503 

NULL Deviance                : 0.39703

Fit Deviance                 : 0.20918

Explained Deviance           : 0.18785

Percent Deviance Explained   : 47.313

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.05 

Confusion Matrix: 

         observed

predicted    1    0

        1  207  520

        0   26 3907

AUC                          :  0.9427 

Percent Correctly Classified :  88.28326 

Sensitivity                  :  0.888412 

Specificity                  :  0.882539 

Kappa                        :  0.3846501 

True Skill Statistic         :  0.770951 

   Calibration Statistics 

Intercept (general calibration)                            :  0.000399389 

Slope   (direction and variation in fit)                   :  1.000503 

Testa0b1 (overall reliability of predictors)               :  0.9999411 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9968319 

Testb1|a (refinement given correct calibration)            :  0.9919399 

============================================================ 

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6049815  (sd 0.07336) 

NULL Deviance                : 0.39701 (sd 0.005652)

Fit Deviance                 : 0.221 (sd 0.031722)

Explained Deviance           : 0.17601 (sd 0.032294)

Percent Deviance Explained   : 44.323 (sd 8.0767)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.052 (sd 0.0042164) 

Confusion Matrix: 

         observed

predicted    1    0

        1  202  530

        0   31 3897

AUC                          :  0.93272  (sd 0.024162) 

Percent Correctly Classified :  87.9612  (sd 1.1836) 

Sensitivity                  :  0.8668478  (sd 0.062914) 

Specificity                  :  0.8802815  (sd 0.012984) 

Kappa                        :  0.3717468  (sd 0.032706) 

True Skill Statistic         :  0.7471293  (sd 0.060622) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.03798806  (sd 0.33576) 

Slope   (direction and variation in fit)                   :  0.9780029  (sd 0.16708) 

Testa0b1 (overall reliability of predictors)               :  0.5878709  (sd 0.29351) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.6098764  (sd 0.18539) 

Testb1|a (refinement given correct calibration)            :  0.4481191  (sd 0.27505) 

Total time = 2.52 min

Boosted Regression Tree Modeling Results

Data:

C:\Withrow\Workspace\Model\brt_1\CovariateCorrelationOutputMDS_initial.csv

n(pres)                 =      233

n(abs)                  =      4427

n covariates considered =      5

total time for model fitting = 0.17min

Settings:

random seed used             : 19680828

tree complexity              : 8

learning rate                : 0.0718

n(trees)                     : 300

model simplification         : cross-validation

n folds                      : 3

n covariates in final model  : 4

Relative influence of predictors in final model:

          Var  rel.inf

        bio16 36.16101

        bio02 24.47096

        bio01 21.46669

 di_all_model 17.90133

Important interactions in final model:

 v1 name1 v2 name2

  3 bio16  2 bio02

  3 bio16  1 bio01

============================================================ 

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.8506478 

NULL Deviance                : 0.39703

Fit Deviance                 : 0.10444

Explained Deviance           : 0.29259

Percent Deviance Explained   : 73.694

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.09 

Confusion Matrix: 

         observed

predicted    1    0

        1  222  212

        0   11 4215

AUC                          :  0.9916 

Percent Correctly Classified :  95.21459 

Sensitivity                  :  0.9527897 

Specificity                  :  0.952112 

Kappa                        :  0.642399 

True Skill Statistic         :  0.9049017 

   Calibration Statistics 

Intercept (general calibration)                            :  0.6888482 

Slope   (direction and variation in fit)                   :  1.576671 

Testa0b1 (overall reliability of predictors)               :  3.164136e-14 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9063842 

Testb1|a (refinement given correct calibration)            :  3.219647e-15 

============================================================ 

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.5517907  (sd 0.067317) 

NULL Deviance                : 0.39701 (sd 0.005652)

Fit Deviance                 : 0.24456 (sd 0.030863)

Explained Deviance           : 0.15244 (sd 0.030084)

Percent Deviance Explained   : 38.41 (sd 7.6612)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.092 (sd 0.0063246) 

Confusion Matrix: 

         observed

predicted    1    0

        1  168  253

        0   65 4174

AUC                          :  0.93724  (sd 0.013488) 

Percent Correctly Classified :  93.17544  (sd 0.92391) 

Sensitivity                  :  0.7210145  (sd 0.091102) 

Specificity                  :  0.9428526  (sd 0.012033) 

Kappa                        :  0.4807245  (sd 0.047865) 

True Skill Statistic         :  0.6638671  (sd 0.084308) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.2873575  (sd 0.30371) 

Slope   (direction and variation in fit)                   :  0.7808831  (sd 0.10644) 

Testa0b1 (overall reliability of predictors)               :  0.1974776  (sd 0.2995) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5626288  (sd 0.35114) 

Testb1|a (refinement given correct calibration)            :  0.1885523  (sd 0.25407) 

Total time = 2.42 min

MARS

Random Forests

MARS Model Results

Data:

C:\Withrow\Workspace\Model\mars_1\CovariateCorrelationOutputMDS_initial.csv

n(pres)                 =      233

n(abs)                  =      4427

n covariates considered =      5

total time for model fitting = 0.04min

Settings:

random seed used             : 123

mars degree                  : 1

mars penalty                 : 2

Summary of Model:

             nsubsets   gcv    rss

bio16              13 100.0  100.0

bio01              12  75.8   76.6

bio02              11  72.0   72.8

di_all_model        8  49.1   50.4

bio15               6  38.9   40.2

============================================================ 

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6321587 

NULL Deviance                : 0.39703

Fit Deviance                 : 0.20754

Explained Deviance           : 0.18949

Percent Deviance Explained   : 47.726

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.06 

Confusion Matrix: 

         observed

predicted    1    0

        1  204  502

        0   29 3925

AUC                          :  0.9438 

Percent Correctly Classified :  88.60515 

Sensitivity                  :  0.8755365 

Specificity                  :  0.8866049 

Kappa                        :  0.3885306 

True Skill Statistic         :  0.7621414 

   Calibration Statistics 

Intercept (general calibration)                            :  0.0007413135 

Slope   (direction and variation in fit)                   :  1.000751 

Testa0b1 (overall reliability of predictors)               :  0.9998868 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9968348 

Testb1|a (refinement given correct calibration)            :  0.9884184 

============================================================ 

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6191279  (sd 0.063331) 

NULL Deviance                : 0.39701 (sd 0.005652)

Fit Deviance                 : 0.21553 (sd 0.023887)

Explained Deviance           : 0.18148 (sd 0.023783)

Percent Deviance Explained   : 45.712 (sd 6.0059)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.053 (sd 0.0048305) 

Confusion Matrix: 

         observed

predicted    1    0

        1  202  570

        0   31 3857

AUC                          :  0.9377  (sd 0.015375) 

Percent Correctly Classified :  87.10219  (sd 0.94679) 

Sensitivity                  :  0.8668478  (sd 0.062914) 

Specificity                  :  0.8712435  (sd 0.01055) 

Kappa                        :  0.3525133  (sd 0.029477) 

True Skill Statistic         :  0.7380913  (sd 0.060224) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.02219644  (sd 0.26816) 

Slope   (direction and variation in fit)                   :  0.9867053  (sd 0.1372) 

Testa0b1 (overall reliability of predictors)               :  0.6276483  (sd 0.24436) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5492176  (sd 0.20854) 

Testb1|a (refinement given correct calibration)            :  0.5187132  (sd 0.22096) 

Total time = 0.43 min

Random Forest Modeling Results

Data:

C:\Withrow\Workspace\Model\rf_1\CovariateCorrelationOutputMDS_initial.csv

n(pres)                 =      233

n(abs)                  =      4427

n covariates considered =      5

total time for model fitting = 0.1min

Settings:

random seed used                       : 19680828

n covariates considered at each split  : 5

n trees                                : 1000

Relative performance of predictors in final model:

                   0        1 MeanDecreaseAccuracy MeanDecreaseGini

bio16        74.2453 150.5163             118.3171          79.7528

bio02        84.3693  19.7013              88.9577          65.3076

bio15        79.6274  30.3048              87.7276          32.3139

bio01        63.6746  67.1785              75.5653          53.2587

di_all_model 29.4376  48.9342              45.8017          49.2461

============================================================ 

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6065517 

NULL Deviance                : 0.39703

Fit Deviance                 : 0.21751

Explained Deviance           : 0.17952

Percent Deviance Explained   : 45.217

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.04 

Confusion Matrix: 

         observed

predicted    1    0

        1  205  585

        0   28 3842

AUC                          :  0.9377 

Percent Correctly Classified :  86.84549 

Sensitivity                  :  0.8798283 

Specificity                  :  0.8678563 

Kappa                        :  0.3506356 

True Skill Statistic         :  0.7476847 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.2849504 

Slope   (direction and variation in fit)                   :  0.8769871 

Testa0b1 (overall reliability of predictors)               :  0.006218378 

Testa0|b1(incorrect calibration given correct refinement)  :  0.1618617 

Testb1|a (refinement given correct calibration)            :  0.0041804 

============================================================ 

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6142666  (sd 0.04331) 

NULL Deviance                : 0.39701 (sd 0.005652)

Fit Deviance                 : 0.20856 (sd 0.016832)

Explained Deviance           : 0.18845 (sd 0.018425)

Percent Deviance Explained   : 47.451 (sd 4.4064)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.4665 (sd 0.017646) 

Confusion Matrix: 

         observed

predicted    1    0

        1  104   64

        0  129 4363

AUC                          :  0.9513  (sd 0.011837) 

Percent Correctly Classified :  95.85785  (sd 0.72644) 

Sensitivity                  :  0.4461957  (sd 0.10303) 

Specificity                  :  0.9855423  (sd 0.0052398) 

Kappa                        :  0.494351  (sd 0.10186) 

True Skill Statistic         :  0.431738  (sd 0.10348) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.1257495  (sd 0.2799) 

Slope   (direction and variation in fit)                   :  0.9275221  (sd 0.09161) 

Testa0b1 (overall reliability of predictors)               :  0.6054353  (sd 0.27083) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5944788  (sd 0.28381) 

Testb1|a (refinement given correct calibration)            :  0.6006185  (sd 0.3145) 

Total time = 1.3 min