Species 103: Pinus attenuata

General Linear Models

Boosted Regression Trees

Generalized Linear Model Results

Data:

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

n(pres)                 =      153

n(abs)                  =      3907

n covariates considered =      5

total time for model fitting = 0.31min

Settings:

model family          : binomial

simplification method : AIC

Results:

number covariates in final model   : 8

Call:

glm(formula = response ~ I(bio14^2) + bio14 + bio03 + I(bio08^2) + 

    bio08 + I(bio03^2) + bio03:bio08 + bio14:bio08, family = model.family, 

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

Deviance Residuals: 

    Min       1Q   Median       3Q      Max  

-1.1968  -0.2817  -0.1472  -0.0370   4.2199  

Coefficients:

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

(Intercept) 15.4958776 11.2637211   1.376 0.168904    

I(bio14^2)  -0.0205254  0.0066673  -3.078 0.002080 ** 

bio14        0.4898582  0.1644648   2.978 0.002897 ** 

bio03       -1.6160605  0.4776134  -3.384 0.000715 ***

I(bio08^2)  -0.0007006  0.0001935  -3.621 0.000293 ***

bio08        0.5544576  0.0671201   8.261  < 2e-16 ***

I(bio03^2)   0.0250379  0.0052490   4.770 1.84e-06 ***

bio03:bio08 -0.0104152  0.0014698  -7.086 1.38e-12 ***

bio14:bio08 -0.0029344  0.0016959  -1.730 0.083586 .  

---

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

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 1303.4  on 4059  degrees of freedom

Residual deviance: 1008.1  on 4051  degrees of freedom

AIC: 1026.1

Number of Fisher Scoring iterations: 12

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.3222336 

NULL Deviance                : 0.32103

Fit Deviance                 : 0.24829

Explained Deviance           : 0.072739

Percent Deviance Explained   : 22.658

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.04 

Confusion Matrix: 

         observed

predicted    1    0

        1  128  987

        0   25 2920

AUC                          :  0.8655 

Percent Correctly Classified :  75.07389 

Sensitivity                  :  0.8366013 

Specificity                  :  0.7473765 

Kappa                        :  0.1452436 

True Skill Statistic         :  0.5839778 

   Calibration Statistics 

Intercept (general calibration)                            :  0.002144241 

Slope   (direction and variation in fit)                   :  1.001046 

Testa0b1 (overall reliability of predictors)               :  0.9999018 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9971771 

Testb1|a (refinement given correct calibration)            :  0.9891816 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.3130636  (sd 0.085451) 

NULL Deviance                : 0.32099 (sd 0.0071546)

Fit Deviance                 : 0.25391 (sd 0.037622)

Explained Deviance           : 0.067072 (sd 0.039997)

Percent Deviance Explained   : 20.804 (sd 12.346)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.045 (sd 0.0052705) 

Confusion Matrix: 

         observed

predicted    1    0

        1  120  848

        0   33 3059

AUC                          :  0.86528  (sd 0.051902) 

Percent Correctly Classified :  78.29284  (sd 3.4715) 

Sensitivity                  :  0.7833333  (sd 0.12615) 

Specificity                  :  0.7828788  (sd 0.034901) 

Kappa                        :  0.1622222  (sd 0.04886) 

True Skill Statistic         :  0.5662121  (sd 0.13665) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.05831788  (sd 0.83576) 

Slope   (direction and variation in fit)                   :  0.9967887  (sd 0.32625) 

Testa0b1 (overall reliability of predictors)               :  0.565118  (sd 0.33633) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.7947357  (sd 0.12585) 

Testb1|a (refinement given correct calibration)            :  0.3940781  (sd 0.30469) 

Total time = 4.18 min

Boosted Regression Tree Modeling Results

Data:

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

n(pres)                 =      153

n(abs)                  =      3907

n covariates considered =      5

total time for model fitting = 0.61min

Settings:

random seed used             : 19680828

tree complexity              : 8

learning rate                : 0.0836

n(trees)                     : 300

model simplification         : cross-validation

n folds                      : 3

n covariates in final model  : 2

Relative influence of predictors in final model:

   Var  rel.inf

 bio08 68.46825

 bio03 31.53175

Important interactions in final model:

 v1 name1 v2 name2

  2 bio08  1 bio03

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.5653624 

NULL Deviance                : 0.32103

Fit Deviance                 : 0.17439

Explained Deviance           : 0.14664

Percent Deviance Explained   : 45.677

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.07 

Confusion Matrix: 

         observed

predicted    1    0

        1  138  380

        0   15 3527

AUC                          :  0.9596 

Percent Correctly Classified :  90.27094 

Sensitivity                  :  0.9019608 

Specificity                  :  0.9027387 

Kappa                        :  0.3749591 

True Skill Statistic         :  0.8046995 

   Calibration Statistics 

Intercept (general calibration)                            :  0.6896092 

Slope   (direction and variation in fit)                   :  1.370618 

Testa0b1 (overall reliability of predictors)               :  3.502005e-06 

Testa0|b1(incorrect calibration given correct refinement)  :  0.916268 

Testb1|a (refinement given correct calibration)            :  5.405883e-07 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.2764438  (sd 0.081089) 

NULL Deviance                : 0.32099 (sd 0.0071546)

Fit Deviance                 : 0.27616 (sd 0.034546)

Explained Deviance           : 0.044831 (sd 0.037594)

Percent Deviance Explained   : 13.857 (sd 11.504)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.066 (sd 0.005164) 

Confusion Matrix: 

         observed

predicted    1    0

        1   81  454

        0   72 3453

AUC                          :  0.82969  (sd 0.05686) 

Percent Correctly Classified :  87.04307  (sd 1.5059) 

Sensitivity                  :  0.5283333  (sd 0.11811) 

Specificity                  :  0.8837815  (sd 0.013335) 

Kappa                        :  0.1886509  (sd 0.058086) 

True Skill Statistic         :  0.4121148  (sd 0.12345) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.7450411  (sd 0.50779) 

Slope   (direction and variation in fit)                   :  0.6841334  (sd 0.17282) 

Testa0b1 (overall reliability of predictors)               :  0.2879905  (sd 0.34745) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.6995281  (sd 0.19248) 

Testb1|a (refinement given correct calibration)            :  0.1870175  (sd 0.27374) 

Total time = 5.08 min

MARS

Random Forests

MARS Model Results

Data:

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

n(pres)                 =      153

n(abs)                  =      3907

n covariates considered =      5

total time for model fitting = 0.09min

Settings:

random seed used             : 123

mars degree                  : 1

mars penalty                 : 2

Summary of Model:

             nsubsets   gcv    rss

bio14              14  97.8   98.9

bio03              13 100.0> 100.0>

di_all_model       10  66.6   70.6

bio02               6  45.5   49.9

bio08               4  28.4   34.5

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.3229541 

NULL Deviance                : 0.32103

Fit Deviance                 : 0.24456

Explained Deviance           : 0.076472

Percent Deviance Explained   : 23.821

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.04 

Confusion Matrix: 

         observed

predicted    1    0

        1  123  856

        0   30 3051

AUC                          :  0.8745 

Percent Correctly Classified :  78.17734 

Sensitivity                  :  0.8039216 

Specificity                  :  0.7809061 

Kappa                        :  0.1627397 

True Skill Statistic         :  0.5848276 

   Calibration Statistics 

Intercept (general calibration)                            :  0.0007556467 

Slope   (direction and variation in fit)                   :  1.000462 

Testa0b1 (overall reliability of predictors)               :  0.9999724 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9971546 

Testb1|a (refinement given correct calibration)            :  0.9947969 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.2809671  (sd 0.11227) 

NULL Deviance                : 0.32099 (sd 0.0071546)

Fit Deviance                 : 0.26965 (sd 0.067594)

Explained Deviance           : 0.051341 (sd 0.069063)

Percent Deviance Explained   : 15.895 (sd 21.588)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.043 (sd 0.0048305) 

Confusion Matrix: 

         observed

predicted    1    0

        1  117  833

        0   36 3074

AUC                          :  0.85841  (sd 0.057369) 

Percent Correctly Classified :  78.59486  (sd 3.1973) 

Sensitivity                  :  0.7629167  (sd 0.13876) 

Specificity                  :  0.7867945  (sd 0.033431) 

Kappa                        :  0.1595847  (sd 0.052054) 

True Skill Statistic         :  0.5497112  (sd 0.13843) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.1452222  (sd 0.92609) 

Slope   (direction and variation in fit)                   :  0.9577317  (sd 0.3567) 

Testa0b1 (overall reliability of predictors)               :  0.6043992  (sd 0.40404) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.7993964  (sd 0.17484) 

Testb1|a (refinement given correct calibration)            :  0.50144  (sd 0.37719) 

Total time = 0.83 min

Random Forest Modeling Results

Data:

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

n(pres)                 =      153

n(abs)                  =      3907

n covariates considered =      5

total time for model fitting = 0.2min

Settings:

random seed used                       : 19680828

n covariates considered at each split  : 2

n trees                                : 1000

Relative performance of predictors in final model:

                   0       1 MeanDecreaseAccuracy MeanDecreaseGini

bio08        43.8376 37.0365              48.7077          50.1359

bio03        41.3548 66.0898              47.7084          30.4826

bio14        35.6197 38.5056              42.4204          27.6457

di_all_model 30.4478 38.0975              40.9306          31.8594

bio02        32.7213 12.1464              36.9986          40.6442

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.4737549 

NULL Deviance                : 0.32103

Fit Deviance                 : 0.21733

Explained Deviance           : 0.1037

Percent Deviance Explained   : 32.303

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.04 

Confusion Matrix: 

         observed

predicted    1    0

        1  124  682

        0   29 3225

AUC                          :  0.8942 

Percent Correctly Classified :  82.48768 

Sensitivity                  :  0.8104575 

Specificity                  :  0.8254415 

Kappa                        :  0.2084628 

True Skill Statistic         :  0.635899 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.2427658 

Slope   (direction and variation in fit)                   :  0.9095443 

Testa0b1 (overall reliability of predictors)               :  0.2162905 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5182278 

Testb1|a (refinement given correct calibration)            :  0.1038862 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.4617788  (sd 0.10034) 

NULL Deviance                : 0.32099 (sd 0.0071546)

Fit Deviance                 : 0.22195 (sd 0.028307)

Explained Deviance           : 0.099033 (sd 0.031567)

Percent Deviance Explained   : 30.761 (sd 9.488)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.357 (sd 0.045043) 

Confusion Matrix: 

         observed

predicted    1    0

        1   43   56

        0  110 3851

AUC                          :  0.89299  (sd 0.035624) 

Percent Correctly Classified :  95.91054  (sd 0.78733) 

Sensitivity                  :  0.2804167  (sd 0.096846) 

Specificity                  :  0.9856639  (sd 0.0073879) 

Kappa                        :  0.3182706  (sd 0.10581) 

True Skill Statistic         :  0.2660805  (sd 0.096664) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.3424691  (sd 0.4829) 

Slope   (direction and variation in fit)                   :  0.8616272  (sd 0.19032) 

Testa0b1 (overall reliability of predictors)               :  0.4240638  (sd 0.29621) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5869922  (sd 0.27244) 

Testb1|a (refinement given correct calibration)            :  0.3959211  (sd 0.33633) 

Total time = 2.74 min