Species 020: Abies magnifica

General Linear Models

Boosted Regression Trees

Generalized Linear Model Results

Data:

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

n(pres)                 =      195

n(abs)                  =      4807

n covariates considered =      7

total time for model fitting = 0.66min

Settings:

model family          : binomial

simplification method : AIC

Results:

number covariates in final model   : 15

Call:

glm(formula = response ~ I(bio10^2) + I(bio14^2) + bio09 + bio15 + 

    I(bio03^2) + bio03 + bio14 + di_all_model + I(bio09^2) + 

    bio10 + I(bio15^2) + bio03:di_all_model + bio14:di_all_model + 

    di_all_model:bio10 + bio15:bio14, family = model.family, 

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

Deviance Residuals: 

    Min       1Q   Median       3Q      Max  

-2.1768  -0.0041   0.0000   0.0000   4.4658  

Coefficients:

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

(Intercept)        -2.627e+02  5.315e+01  -4.943 7.69e-07 ***

I(bio10^2)         -3.421e-03  9.058e-04  -3.777 0.000159 ***

I(bio14^2)         -5.703e-02  1.286e-02  -4.434 9.25e-06 ***

bio09              -3.670e-01  1.151e-01  -3.187 0.001439 ** 

bio15              -8.153e-01  3.024e-01  -2.696 0.007018 ** 

I(bio03^2)         -1.342e-01  2.647e-02  -5.068 4.03e-07 ***

bio03               1.208e+01  2.394e+00   5.047 4.48e-07 ***

bio14               3.016e-01  8.782e-01   0.343 0.731237    

di_all_model        2.309e+00  3.752e-01   6.154 7.54e-10 ***

I(bio09^2)          2.716e-03  7.993e-04   3.398 0.000679 ***

bio10               5.117e-01  1.536e-01   3.331 0.000864 ***

I(bio15^2)          6.801e-03  2.016e-03   3.373 0.000743 ***

bio03:di_all_model -2.489e-02  6.091e-03  -4.086 4.39e-05 ***

bio14:di_all_model -3.944e-02  6.508e-03  -6.060 1.36e-09 ***

di_all_model:bio10 -5.890e-03  1.098e-03  -5.365 8.11e-08 ***

bio15:bio14         1.835e-02  9.732e-03   1.886 0.059332 .  

---

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

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 1647.69  on 5001  degrees of freedom

Residual deviance:  429.52  on 4986  degrees of freedom

AIC: 461.52

Number of Fisher Scoring iterations: 14

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.8075041 

NULL Deviance                : 0.32941

Fit Deviance                 : 0.085871

Explained Deviance           : 0.24354

Percent Deviance Explained   : 73.932

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.07 

Confusion Matrix: 

         observed

predicted    1    0

        1  188  188

        0    7 4619

AUC                          :  0.9917 

Percent Correctly Classified :  96.10156 

Sensitivity                  :  0.9641026 

Specificity                  :  0.9608904 

Kappa                        :  0.6400113 

True Skill Statistic         :  0.9249929 

   Calibration Statistics 

Intercept (general calibration)                            :  0.0008175833 

Slope   (direction and variation in fit)                   :  1.003496 

Testa0b1 (overall reliability of predictors)               :  0.9988528 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9951044 

Testb1|a (refinement given correct calibration)            :  0.9620988 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.7933885  (sd 0.047822) 

NULL Deviance                : 0.32938 (sd 0.0065522)

Fit Deviance                 : 0.095042 (sd 0.025135)

Explained Deviance           : 0.23433 (sd 0.026049)

Percent Deviance Explained   : 71.133 (sd 7.785)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.075 (sd 0.0097183) 

Confusion Matrix: 

         observed

predicted    1    0

        1  182  185

        0   13 4622

AUC                          :  0.98979  (sd 0.0057616) 

Percent Correctly Classified :  96.04215  (sd 0.81743) 

Sensitivity                  :  0.9334211  (sd 0.054725) 

Specificity                  :  0.9615172  (sd 0.0088225) 

Kappa                        :  0.6316821  (sd 0.050381) 

True Skill Statistic         :  0.8949383  (sd 0.053009) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.004901783  (sd 0.2887) 

Slope   (direction and variation in fit)                   :  0.9690362  (sd 0.28837) 

Testa0b1 (overall reliability of predictors)               :  0.5188681  (sd 0.29879) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5930299  (sd 0.22795) 

Testb1|a (refinement given correct calibration)            :  0.4168679  (sd 0.30081) 

Total time = 6.53 min

Boosted Regression Tree Modeling Results

Data:

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

n(pres)                 =      195

n(abs)                  =      4807

n covariates considered =      7

total time for model fitting = 0.32min

Settings:

random seed used             : 19680828

tree complexity              : 9

learning rate                : 0.0811

n(trees)                     : 300

model simplification         : cross-validation

n folds                      : 3

n covariates in final model  : 5

Relative influence of predictors in final model:

   Var   rel.inf

 bio10 28.856361

 bio12 23.473090

 bio15 20.295401

 bio14 18.701929

 bio03  8.673219

Important interactions in final model:

 v1 name1 v2 name2

  4 bio14  2 bio10

  3 bio12  2 bio10

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.9888062 

NULL Deviance                : 0.32941

Fit Deviance                 : 0.01304

Explained Deviance           : 0.31637

Percent Deviance Explained   : 96.041

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.385 

Confusion Matrix: 

         observed

predicted    1    0

        1  195    0

        0    0 4807

AUC                          :  1 

Percent Correctly Classified :  100 

Sensitivity                  :  1 

Specificity                  :  1 

Kappa                        :  1 

True Skill Statistic         :  1 

   Calibration Statistics 

Intercept (general calibration)                            :  29.90151 

Slope   (direction and variation in fit)                   :  62.89941 

Testa0b1 (overall reliability of predictors)               :  6.550316e-15 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9836434 

Testb1|a (refinement given correct calibration)            :  6.661338e-16 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.7358395  (sd 0.074194) 

NULL Deviance                : 0.32938 (sd 0.0065522)

Fit Deviance                 : 0.12623 (sd 0.032076)

Explained Deviance           : 0.20314 (sd 0.035756)

Percent Deviance Explained   : 61.576 (sd 10.124)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.3265 (sd 0.10223) 

Confusion Matrix: 

         observed

predicted    1    0

        1  136   75

        0   59 4732

AUC                          :  0.98535  (sd 0.0070939) 

Percent Correctly Classified :  97.32119  (sd 1.1554) 

Sensitivity                  :  0.6957895  (sd 0.12435) 

Specificity                  :  0.9844027  (sd 0.011561) 

Kappa                        :  0.6595768  (sd 0.11496) 

True Skill Statistic         :  0.6801922  (sd 0.12369) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.07101521  (sd 0.40546) 

Slope   (direction and variation in fit)                   :  0.7518502  (sd 0.10413) 

Testa0b1 (overall reliability of predictors)               :  0.2274844  (sd 0.3005) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.4562794  (sd 0.28891) 

Testb1|a (refinement given correct calibration)            :  0.1605794  (sd 0.23741) 

Total time = 3.1 min

MARS

Random Forests

MARS Model Results

Data:

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

n(pres)                 =      195

n(abs)                  =      4807

n covariates considered =      7

total time for model fitting = 0.05min

Settings:

random seed used             : 123

mars degree                  : 1

mars penalty                 : 2

Summary of Model:

             nsubsets   gcv    rss

bio10              16 100.0  100.0

bio14              14  72.9   73.5

di_all_model       13  54.6   55.8

bio15              11  46.8   48.1

bio03               8  30.1   31.7

bio12               5  17.8   19.7

bio09               3   9.7   11.9

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.7690535 

NULL Deviance                : 0.32941

Fit Deviance                 : 0.099552

Explained Deviance           : 0.22985

Percent Deviance Explained   : 69.778

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.09 

Confusion Matrix: 

         observed

predicted    1    0

        1  187  199

        0    8 4608

AUC                          :  0.9888 

Percent Correctly Classified :  95.86166 

Sensitivity                  :  0.9589744 

Specificity                  :  0.958602 

Kappa                        :  0.624254 

True Skill Statistic         :  0.9175764 

   Calibration Statistics 

Intercept (general calibration)                            :  0.0007133962 

Slope   (direction and variation in fit)                   :  1.001847 

Testa0b1 (overall reliability of predictors)               :  0.9995877 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9953809 

Testb1|a (refinement given correct calibration)            :  0.9775579 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.7460489  (sd 0.034581) 

NULL Deviance                : 0.32938 (sd 0.0065522)

Fit Deviance                 : 0.11217 (sd 0.017382)

Explained Deviance           : 0.2172 (sd 0.018442)

Percent Deviance Explained   : 65.934 (sd 5.3301)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.085 (sd 0.01354) 

Confusion Matrix: 

         observed

predicted    1    0

        1  183  227

        0   12 4580

AUC                          :  0.985  (sd 0.0067731) 

Percent Correctly Classified :  95.22267  (sd 1.4656) 

Sensitivity                  :  0.9384211  (sd 0.039779) 

Specificity                  :  0.9527858  (sd 0.015478) 

Kappa                        :  0.5913155  (sd 0.080896) 

True Skill Statistic         :  0.8912069  (sd 0.039593) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.03967198  (sd 0.37004) 

Slope   (direction and variation in fit)                   :  0.9489021  (sd 0.13885) 

Testa0b1 (overall reliability of predictors)               :  0.5659838  (sd 0.27488) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5169641  (sd 0.25051) 

Testb1|a (refinement given correct calibration)            :  0.5449553  (sd 0.24789) 

Total time = 0.59 min

Random Forest Modeling Results

Data:

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

n(pres)                 =      195

n(abs)                  =      4807

n covariates considered =      7

total time for model fitting = 0.11min

Settings:

random seed used                       : 19680828

n covariates considered at each split  : 3

n trees                                : 1000

Relative performance of predictors in final model:

                   0       1 MeanDecreaseAccuracy MeanDecreaseGini

bio10        44.0280 75.7081              63.3723          46.8452

bio03        51.9742 23.8794              59.1569          22.3452

bio14        42.5782 39.5556              48.5770          39.5446

di_all_model 24.6222 38.3208              40.0840          23.5777

bio09        30.6303 29.6398              37.2735          25.5106

bio12        21.5290 50.6364              36.4988          40.8956

bio15        27.9912 37.6882              32.3307          37.8963

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.8014537 

NULL Deviance                : 0.32941

Fit Deviance                 : 0.097234

Explained Deviance           : 0.23217

Percent Deviance Explained   : 70.482

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.05 

Confusion Matrix: 

         observed

predicted    1    0

        1  184  268

        0   11 4539

AUC                          :  0.9867 

Percent Correctly Classified :  94.42223 

Sensitivity                  :  0.9435897 

Specificity                  :  0.944248 

Kappa                        :  0.5439374 

True Skill Statistic         :  0.8878377 

   Calibration Statistics 

Intercept (general calibration)                            :  0.07814332 

Slope   (direction and variation in fit)                   :  1.181854 

Testa0b1 (overall reliability of predictors)               :  0.004401969 

Testa0|b1(incorrect calibration given correct refinement)  :  0.2693134 

Testb1|a (refinement given correct calibration)            :  0.001913031 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.7973668  (sd 0.047186) 

NULL Deviance                : 0.32938 (sd 0.0065522)

Fit Deviance                 : 0.096543 (sd 0.021529)

Explained Deviance           : 0.23283 (sd 0.024238)

Percent Deviance Explained   : 70.645 (sd 6.7839)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.477 (sd 0.0078881) 

Confusion Matrix: 

         observed

predicted    1    0

        1  136   29

        0   59 4778

AUC                          :  0.98867  (sd 0.0070616) 

Percent Correctly Classified :  98.24076  (sd 0.48762) 

Sensitivity                  :  0.6957895  (sd 0.08707) 

Specificity                  :  0.9939674  (sd 0.0035939) 

Kappa                        :  0.7448246  (sd 0.071282) 

True Skill Statistic         :  0.6897569  (sd 0.087318) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.2421894  (sd 0.45577) 

Slope   (direction and variation in fit)                   :  1.257461  (sd 0.25856) 

Testa0b1 (overall reliability of predictors)               :  0.4425566  (sd 0.3267) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5892847  (sd 0.27761) 

Testb1|a (refinement given correct calibration)            :  0.336022  (sd 0.27525) 

Total time = 1.53 min