Species 021: Abies xshastensis

General Linear Models

Boosted Regression Trees

Generalized Linear Model Results

Data:

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

n(pres)                 =      107

n(abs)                  =      2201

n covariates considered =      6

total time for model fitting = 0.25min

Settings:

model family          : binomial

simplification method : AIC

Results:

number covariates in final model   : 18

Call:

glm(formula = response ~ I(bio10^2) + bio10 + bio03 + I(bio03^2) + 

    bio15 + bio17 + I(bio17^2) + I(bio04^2) + bio04 + bio10:bio17 + 

    bio10:bio04 + bio10:bio15 + bio03:bio17 + bio10:bio03 + bio17:bio04 + 

    bio15:bio17 + bio03:bio04 + bio03:bio15, family = model.family, 

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

Deviance Residuals: 

     Min        1Q    Median        3Q       Max  

-1.94596  -0.08669  -0.00661  -0.00002   3.07124  

Coefficients:

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

(Intercept) -7.854e+02  3.947e+02  -1.990 0.046601 *  

I(bio10^2)  -8.678e-03  1.413e-03  -6.140 8.26e-10 ***

bio10       -5.128e+00  1.297e+00  -3.955 7.65e-05 ***

bio03        3.868e+01  1.301e+01   2.974 0.002942 ** 

I(bio03^2)  -3.484e-01  1.017e-01  -3.425 0.000615 ***

bio15        1.330e+00  1.448e+00   0.919 0.358292    

bio17       -4.329e+00  9.808e-01  -4.414 1.01e-05 ***

I(bio17^2)  -3.746e-03  7.921e-04  -4.729 2.26e-06 ***

I(bio04^2)  -7.616e-06  2.146e-06  -3.549 0.000387 ***

bio04        1.249e-01  5.855e-02   2.134 0.032872 *  

bio10:bio17 -8.122e-03  1.702e-03  -4.771 1.83e-06 ***

bio10:bio04  5.158e-04  1.004e-04   5.137 2.80e-07 ***

bio10:bio15  1.661e-02  3.181e-03   5.221 1.78e-07 ***

bio03:bio17  6.728e-02  1.610e-02   4.179 2.93e-05 ***

bio10:bio03  8.624e-02  2.212e-02   3.898 9.69e-05 ***

bio17:bio04  3.725e-04  8.036e-05   4.635 3.57e-06 ***

bio15:bio17  1.217e-02  3.942e-03   3.087 0.002019 ** 

bio03:bio04 -2.942e-03  1.021e-03  -2.881 0.003967 ** 

bio03:bio15 -9.539e-02  3.529e-02  -2.703 0.006865 ** 

---

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

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 866.22  on 2307  degrees of freedom

Residual deviance: 392.87  on 2289  degrees of freedom

AIC: 430.87

Number of Fisher Scoring iterations: 13

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6631697 

NULL Deviance                : 0.37531

Fit Deviance                 : 0.17022

Explained Deviance           : 0.20509

Percent Deviance Explained   : 54.645

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.07 

Confusion Matrix: 

         observed

predicted    1    0

        1   97  216

        0   10 1985

AUC                          :  0.9664 

Percent Correctly Classified :  90.20797 

Sensitivity                  :  0.9065421 

Specificity                  :  0.9018628 

Kappa                        :  0.4219629 

True Skill Statistic         :  0.8084048 

   Calibration Statistics 

Intercept (general calibration)                            :  0.0009203704 

Slope   (direction and variation in fit)                   :  1.001178 

Testa0b1 (overall reliability of predictors)               :  0.9999055 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9976111 

Testb1|a (refinement given correct calibration)            :  0.9892951 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6102008  (sd 0.10901) 

NULL Deviance                : 0.37521 (sd 0.012364)

Fit Deviance                 : 0.19524 (sd 0.045225)

Explained Deviance           : 0.17997 (sd 0.04489)

Percent Deviance Explained   : 47.972 (sd 11.963)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.07 (sd 0.004714) 

Confusion Matrix: 

         observed

predicted    1    0

        1   92  221

        0   15 1980

AUC                          :  0.95492  (sd 0.020735) 

Percent Correctly Classified :  89.7764  (sd 2.5132) 

Sensitivity                  :  0.8609091  (sd 0.12362) 

Specificity                  :  0.8995928  (sd 0.027637) 

Kappa                        :  0.4026894  (sd 0.075746) 

True Skill Statistic         :  0.7605019  (sd 0.11792) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.02730155  (sd 0.53802) 

Slope   (direction and variation in fit)                   :  1.072086  (sd 0.65418) 

Testa0b1 (overall reliability of predictors)               :  0.3982633  (sd 0.32592) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.7324866  (sd 0.20958) 

Testb1|a (refinement given correct calibration)            :  0.2368648  (sd 0.22529) 

Total time = 2.07 min

Boosted Regression Tree Modeling Results

Data:

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

n(pres)                 =      107

n(abs)                  =      2201

n covariates considered =      6

total time for model fitting = 0.17min

Settings:

random seed used             : 19680828

tree complexity              : 5

learning rate                : 0.0364

n(trees)                     : 300

model simplification         : cross-validation

n folds                      : 3

n covariates in final model  : 3

Relative influence of predictors in final model:

   Var  rel.inf

 bio10 36.30852

 bio17 32.78556

 bio04 30.90592

Important interactions in final model:

 v1 name1 v2 name2

  3 bio17  2 bio10

  2 bio10  1 bio04

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.7910162 

NULL Deviance                : 0.37531

Fit Deviance                 : 0.12674

Explained Deviance           : 0.24857

Percent Deviance Explained   : 66.231

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.1 

Confusion Matrix: 

         observed

predicted    1    0

        1  101  127

        0    6 2074

AUC                          :  0.9877 

Percent Correctly Classified :  94.23744 

Sensitivity                  :  0.9439252 

Specificity                  :  0.942299 

Kappa                        :  0.5762437 

True Skill Statistic         :  0.8862242 

   Calibration Statistics 

Intercept (general calibration)                            :  0.9159794 

Slope   (direction and variation in fit)                   :  1.72324 

Testa0b1 (overall reliability of predictors)               :  9.50934e-10 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9698975 

Testb1|a (refinement given correct calibration)            :  1.151458e-10 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.4716633  (sd 0.14868) 

NULL Deviance                : 0.37521 (sd 0.012364)

Fit Deviance                 : 0.25804 (sd 0.055902)

Explained Deviance           : 0.11717 (sd 0.057029)

Percent Deviance Explained   : 31.167 (sd 15.019)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.078 (sd 0.03084) 

Confusion Matrix: 

         observed

predicted    1    0

        1   77  220

        0   30 1981

AUC                          :  0.89863  (sd 0.068767) 

Percent Correctly Classified :  89.16582  (sd 3.1202) 

Sensitivity                  :  0.7236364  (sd 0.13974) 

Specificity                  :  0.9000494  (sd 0.034203) 

Kappa                        :  0.3472521  (sd 0.091085) 

True Skill Statistic         :  0.6236857  (sd 0.1335) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.05057701  (sd 0.86714) 

Slope   (direction and variation in fit)                   :  0.9627926  (sd 0.37778) 

Testa0b1 (overall reliability of predictors)               :  0.3508631  (sd 0.29143) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.6377932  (sd 0.23535) 

Testb1|a (refinement given correct calibration)            :  0.2160722  (sd 0.19153) 

Total time = 1.73 min

MARS

Random Forests

MARS Model Results

Data:

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

n(pres)                 =      107

n(abs)                  =      2201

n covariates considered =      6

total time for model fitting = 0.03min

Settings:

random seed used             : 123

mars degree                  : 1

mars penalty                 : 2

Summary of Model:

      nsubsets   gcv    rss

bio10       13 100.0  100.0

bio03       11  76.5   78.1

bio17       10  57.8   61.5

bio15        9  55.2   58.7

bio04        8  50.2   53.7

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.5404035 

NULL Deviance                : 0.37531

Fit Deviance                 : 0.21602

Explained Deviance           : 0.1593

Percent Deviance Explained   : 42.444

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.06 

Confusion Matrix: 

         observed

predicted    1    0

        1   92  312

        0   15 1889

AUC                          :  0.9397 

Percent Correctly Classified :  85.83189 

Sensitivity                  :  0.8598131 

Specificity                  :  0.8582463 

Kappa                        :  0.3094575 

True Skill Statistic         :  0.7180593 

   Calibration Statistics 

Intercept (general calibration)                            :  0.001087479 

Slope   (direction and variation in fit)                   :  1.00094 

Testa0b1 (overall reliability of predictors)               :  0.9999369 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9978453 

Testb1|a (refinement given correct calibration)            :  0.9912993 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.4943255  (sd 0.13764) 

NULL Deviance                : 0.37521 (sd 0.012364)

Fit Deviance                 : 0.23229 (sd 0.043588)

Explained Deviance           : 0.14292 (sd 0.044417)

Percent Deviance Explained   : 38.053 (sd 11.71)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.06 (sd 0.0066667) 

Confusion Matrix: 

         observed

predicted    1    0

        1   89  330

        0   18 1871

AUC                          :  0.92636  (sd 0.034254) 

Percent Correctly Classified :  84.92415  (sd 2.8399) 

Sensitivity                  :  0.8354545  (sd 0.14674) 

Specificity                  :  0.850072  (sd 0.025144) 

Kappa                        :  0.2907202  (sd 0.081185) 

True Skill Statistic         :  0.6855265  (sd 0.16225) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.1043819  (sd 0.51311) 

Slope   (direction and variation in fit)                   :  0.986667  (sd 0.38236) 

Testa0b1 (overall reliability of predictors)               :  0.5294701  (sd 0.29365) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.8191362  (sd 0.1644) 

Testb1|a (refinement given correct calibration)            :  0.3414911  (sd 0.26221) 

Total time = 0.3 min

Random Forest Modeling Results

Data:

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

n(pres)                 =      107

n(abs)                  =      2201

n covariates considered =      6

total time for model fitting = 0.07min

Settings:

random seed used                       : 19680828

n covariates considered at each split  : 1

n trees                                : 1000

Relative performance of predictors in final model:

                   0       1 MeanDecreaseAccuracy MeanDecreaseGini

bio03        41.0164 22.9883              45.0781          13.5744

bio10        27.0290 51.4809              42.3034          25.1181

bio17        29.6278 44.4631              39.1323          22.6135

bio15        26.4065 31.0258              34.4718          16.4590

bio04        25.3663 26.8861              32.7169          21.9083

di_all_model 29.7983  4.5706              30.8369          13.0811

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.5794988 

NULL Deviance                : 0.37531

Fit Deviance                 : 0.21549

Explained Deviance           : 0.15983

Percent Deviance Explained   : 42.585

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.06 

Confusion Matrix: 

         observed

predicted    1    0

        1   93  267

        0   14 1934

AUC                          :  0.9321 

Percent Correctly Classified :  87.82496 

Sensitivity                  :  0.8691589 

Specificity                  :  0.8786915 

Kappa                        :  0.3519678 

True Skill Statistic         :  0.7478504 

   Calibration Statistics 

Intercept (general calibration)                            :  0.3338055 

Slope   (direction and variation in fit)                   :  1.117483 

Testa0b1 (overall reliability of predictors)               :  0.1886325 

Testa0|b1(incorrect calibration given correct refinement)  :  0.3142768 

Testb1|a (refinement given correct calibration)            :  0.12745 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.5666863  (sd 0.12496) 

NULL Deviance                : 0.37521 (sd 0.012364)

Fit Deviance                 : 0.21031 (sd 0.036035)

Explained Deviance           : 0.1649 (sd 0.038272)

Percent Deviance Explained   : 43.889 (sd 9.9524)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.466 (sd 0.01022) 

Confusion Matrix: 

         observed

predicted    1    0

        1   37   22

        0   70 2179

AUC                          :  0.93837  (sd 0.033025) 

Percent Correctly Classified :  96.01355  (sd 1.1168) 

Sensitivity                  :  0.3427273  (sd 0.17918) 

Specificity                  :  0.9900041  (sd 0.0059851) 

Kappa                        :  0.4104782  (sd 0.21473) 

True Skill Statistic         :  0.3327314  (sd 0.18145) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.01704954  (sd 0.34417) 

Slope   (direction and variation in fit)                   :  1.037878  (sd 0.23822) 

Testa0b1 (overall reliability of predictors)               :  0.6443415  (sd 0.17945) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.7569441  (sd 0.21961) 

Testb1|a (refinement given correct calibration)            :  0.4543546  (sd 0.20255) 

Total time = 0.79 min