Species 055: Cupressus sargentii

General Linear Models

Boosted Regression Trees

Generalized Linear Model Results

Data:

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

n(pres)                 =      27

n(abs)                  =      1998

n covariates considered =      4

total time for model fitting = 0.15min

Settings:

model family          : binomial

simplification method : AIC

Results:

number covariates in final model   : 6

Call:

glm(formula = response ~ I(bio17^2) + bio17 + bio06 + bio03 + 

    I(bio03^2) + bio06:bio03, family = model.family, data = dat, 

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

Deviance Residuals: 

    Min       1Q   Median       3Q      Max  

-0.7297  -0.1184  -0.0123   0.0000   3.4207  

Coefficients:

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

(Intercept) -4.706e+02  1.711e+02  -2.750  0.00595 ** 

I(bio17^2)  -7.188e-03  1.684e-03  -4.269 1.96e-05 ***

bio17        3.424e-01  8.452e-02   4.051 5.10e-05 ***

bio06       -7.693e-01  3.454e-01  -2.227  0.02592 *  

bio03        1.827e+01  6.769e+00   2.699  0.00695 ** 

I(bio03^2)  -1.796e-01  6.694e-02  -2.683  0.00730 ** 

bio06:bio03  1.473e-02  6.815e-03   2.161  0.03067 *  

---

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

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 286.78  on 2024  degrees of freedom

Residual deviance: 198.22  on 2018  degrees of freedom

AIC: 212.22

Number of Fisher Scoring iterations: 14

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.2647685 

NULL Deviance                : 0.14162

Fit Deviance                 : 0.097884

Explained Deviance           : 0.043737

Percent Deviance Explained   : 30.883

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.04 

Confusion Matrix: 

         observed

predicted    1    0

        1   24  191

        0    3 1807

AUC                          :  0.9315 

Percent Correctly Classified :  90.41975 

Sensitivity                  :  0.8888889 

Specificity                  :  0.9044044 

Kappa                        :  0.1788939 

True Skill Statistic         :  0.7932933 

   Calibration Statistics 

Intercept (general calibration)                            :  0.008644103 

Slope   (direction and variation in fit)                   :  1.003528 

Testa0b1 (overall reliability of predictors)               :  0.9998039 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9967358 

Testb1|a (refinement given correct calibration)            :  0.9845414 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.2317699  (sd 0.089531) 

NULL Deviance                : 0.14117 (sd 0.020667)

Fit Deviance                 : 0.10718 (sd 0.024381)

Explained Deviance           : 0.033989 (sd 0.019267)

Percent Deviance Explained   : 24.085 (sd 14.215)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.035 (sd 0.0052705) 

Confusion Matrix: 

         observed

predicted    1    0

        1   21  211

        0    6 1787

AUC                          :  0.90449  (sd 0.057157) 

Percent Correctly Classified :  89.27646  (sd 1.9571) 

Sensitivity                  :  0.7833333  (sd 0.19325) 

Specificity                  :  0.8943175  (sd 0.019037) 

Kappa                        :  0.1436885  (sd 0.045206) 

True Skill Statistic         :  0.6776508  (sd 0.19926) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.05182028  (sd 1.9152) 

Slope   (direction and variation in fit)                   :  1.069103  (sd 0.83036) 

Testa0b1 (overall reliability of predictors)               :  0.6448076  (sd 0.30704) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.7981151  (sd 0.14847) 

Testb1|a (refinement given correct calibration)            :  0.4742144  (sd 0.33924) 

Total time = 1.45 min

Boosted Regression Tree Modeling Results

Data:

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

n(pres)                 =      27

n(abs)                  =      1998

n covariates considered =      4

total time for model fitting = 0.28min

Settings:

random seed used             : 19680828

tree complexity              : 5

learning rate                : 0.0161

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

 bio17 61.76195

 bio06 38.23805

Important interactions in final model:

 v1 name1 v2 name2

  2 bio17  1 bio06

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.5995383 

NULL Deviance                : 0.14162

Fit Deviance                 : 0.064978

Explained Deviance           : 0.076643

Percent Deviance Explained   : 54.118

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.02 

Confusion Matrix: 

         observed

predicted    1    0

        1   25  151

        0    2 1847

AUC                          :  0.9874 

Percent Correctly Classified :  92.44444 

Sensitivity                  :  0.9259259 

Specificity                  :  0.9244244 

Kappa                        :  0.2284677 

True Skill Statistic         :  0.8503504 

   Calibration Statistics 

Intercept (general calibration)                            :  1.39392 

Slope   (direction and variation in fit)                   :  1.668521 

Testa0b1 (overall reliability of predictors)               :  0.0008130722 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9483714 

Testb1|a (refinement given correct calibration)            :  0.000162185 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.1072695  (sd 0.065244) 

NULL Deviance                : 0.14117 (sd 0.020667)

Fit Deviance                 : 0.12516 (sd 0.024276)

Explained Deviance           : 0.016011 (sd 0.017874)

Percent Deviance Explained   : 11.194 (sd 12.112)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.041 (sd 0.014491) 

Confusion Matrix: 

         observed

predicted    1    0

        1   13  113

        0   14 1885

AUC                          :  0.8731  (sd 0.077218) 

Percent Correctly Classified :  93.72543  (sd 2.201) 

Sensitivity                  :  0.4833333  (sd 0.26586) 

Specificity                  :  0.9433918  (sd 0.023065) 

Kappa                        :  0.1544513  (sd 0.089548) 

True Skill Statistic         :  0.4267251  (sd 0.25882) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.8957377  (sd 0.69083) 

Slope   (direction and variation in fit)                   :  0.7034822  (sd 0.16845) 

Testa0b1 (overall reliability of predictors)               :  0.5776993  (sd 0.27888) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.7127872  (sd 0.10568) 

Testb1|a (refinement given correct calibration)            :  0.4082425  (sd 0.28509) 

Total time = 2.97 min

MARS

Random Forests

MARS Model Results

Data:

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

n(pres)                 =      27

n(abs)                  =      1998

n covariates considered =      4

total time for model fitting = 0.09min

Settings:

random seed used             : 123

mars degree                  : 1

mars penalty                 : 2

Summary of Model:

      nsubsets   gcv    rss

bio17        8 100.0  100.0

bio03        5  66.2   70.1

bio06        3  48.9   52.7

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.2528548 

NULL Deviance                : 0.14162

Fit Deviance                 : 0.099051

Explained Deviance           : 0.04257

Percent Deviance Explained   : 30.059

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.02 

Confusion Matrix: 

         observed

predicted    1    0

        1   24  261

        0    3 1737

AUC                          :  0.9357 

Percent Correctly Classified :  86.96296 

Sensitivity                  :  0.8888889 

Specificity                  :  0.8693694 

Kappa                        :  0.1327201 

True Skill Statistic         :  0.7582583 

   Calibration Statistics 

Intercept (general calibration)                            :  0.003739103 

Slope   (direction and variation in fit)                   :  1.00164 

Testa0b1 (overall reliability of predictors)               :  0.9999258 

Testa0|b1(incorrect calibration given correct refinement)  :  0.996696 

Testb1|a (refinement given correct calibration)            :  0.990858 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.1982072  (sd 0.109) 

NULL Deviance                : 0.14117 (sd 0.020667)

Fit Deviance                 : 0.10944 (sd 0.024539)

Explained Deviance           : 0.031735 (sd 0.019754)

Percent Deviance Explained   : 22.425 (sd 14.42)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.023 (sd 0.0048305) 

Confusion Matrix: 

         observed

predicted    1    0

        1   21  235

        0    6 1763

AUC                          :  0.89973  (sd 0.075725) 

Percent Correctly Classified :  88.09101  (sd 2.2453) 

Sensitivity                  :  0.7666667  (sd 0.26294) 

Specificity                  :  0.8823123  (sd 0.024701) 

Kappa                        :  0.1266794  (sd 0.052968) 

True Skill Statistic         :  0.648979  (sd 0.25123) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.4158698  (sd 1.1691) 

Slope   (direction and variation in fit)                   :  0.8981391  (sd 0.42058) 

Testa0b1 (overall reliability of predictors)               :  0.7101204  (sd 0.29739) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.8018684  (sd 0.16563) 

Testb1|a (refinement given correct calibration)            :  0.5365264  (sd 0.31721) 

Total time = 0.56 min

Random Forest Modeling Results

Data:

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

n(pres)                 =      27

n(abs)                  =      1998

n covariates considered =      4

total time for model fitting = 0.12min

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

bio17        9.5858 15.4889              11.8454           8.1924

bio06        9.1987 13.8891              11.3356           7.9464

bio03        2.9061  9.8395               4.9327           3.9501

di_all_model 0.7223 13.5104               4.2628           5.6753

Model Failed