Species 041: Chamaecyparis lawsoniana

General Linear Models

Boosted Regression Trees

Generalized Linear Model Results

Data:

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

n(pres)                 =      101

n(abs)                  =      1030

n covariates considered =      4

total time for model fitting = 0.05min

Settings:

model family          : binomial

simplification method : AIC

Results:

number covariates in final model   : 5

Call:

glm(formula = response ~ I(bio10^2) + I(bio18^2) + bio18 + I(bio06^2) + 

    bio06, family = model.family, data = dat, weights = weight, 

    na.action = "na.exclude")

Deviance Residuals: 

    Min       1Q   Median       3Q      Max  

-1.6077  -0.2303  -0.0333   0.0000   3.7712  

Coefficients:

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

(Intercept) -2.108e+01  5.062e+00  -4.163 3.14e-05 ***

I(bio10^2)  -1.499e-04  4.519e-05  -3.318 0.000907 ***

I(bio18^2)  -4.027e-03  7.803e-04  -5.160 2.47e-07 ***

bio18        6.335e-01  1.192e-01   5.314 1.07e-07 ***

I(bio06^2)  -1.379e-03  2.256e-04  -6.113 9.80e-10 ***

bio06       -3.493e-02  1.034e-02  -3.378 0.000731 ***

---

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

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 680.68  on 1130  degrees of freedom

Residual deviance: 374.89  on 1125  degrees of freedom

AIC: 386.89

Number of Fisher Scoring iterations: 12

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6386158 

NULL Deviance                : 0.60184

Fit Deviance                 : 0.33147

Explained Deviance           : 0.27037

Percent Deviance Explained   : 44.924

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.15 

Confusion Matrix: 

         observed

predicted   1   0

        1  88 131

        0  13 899

AUC                          :  0.9375 

Percent Correctly Classified :  87.2679 

Sensitivity                  :  0.8712871 

Specificity                  :  0.8728155 

Kappa                        :  0.4873364 

True Skill Statistic         :  0.7441027 

   Calibration Statistics 

Intercept (general calibration)                            :  0.0005685446 

Slope   (direction and variation in fit)                   :  1.000819 

Testa0b1 (overall reliability of predictors)               :  0.9999674 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9988271 

Testb1|a (refinement given correct calibration)            :  0.993661 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6251336  (sd 0.089337) 

NULL Deviance                : 0.60179 (sd 0.012753)

Fit Deviance                 : 0.34559 (sd 0.11017)

Explained Deviance           : 0.2562 (sd 0.10955)

Percent Deviance Explained   : 42.59 (sd 18.248)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.148 (sd 0.0063246) 

Confusion Matrix: 

         observed

predicted   1   0

        1  88 134

        0  13 896

AUC                          :  0.93514  (sd 0.042102) 

Percent Correctly Classified :  86.99984  (sd 3.5129) 

Sensitivity                  :  0.87  (sd 0.12517) 

Specificity                  :  0.8699516  (sd 0.03393) 

Kappa                        :  0.4857119  (sd 0.11076) 

True Skill Statistic         :  0.7399516  (sd 0.13654) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.03221568  (sd 0.44528) 

Slope   (direction and variation in fit)                   :  1.143177  (sd 0.49907) 

Testa0b1 (overall reliability of predictors)               :  0.4819105  (sd 0.34642) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.6231043  (sd 0.24823) 

Testb1|a (refinement given correct calibration)            :  0.4271712  (sd 0.36001) 

Total time = 0.36 min

Boosted Regression Tree Modeling Results

Data:

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

n(pres)                 =      101

n(abs)                  =      1030

n covariates considered =      4

total time for model fitting = 0.15min

Settings:

random seed used             : 19680828

tree complexity              : 4

learning rate                : 0.0073

n(trees)                     : 600

model simplification         : cross-validation

n folds                      : 3

n covariates in final model  : 2

Relative influence of predictors in final model:

   Var  rel.inf

 bio18 66.28442

 bio06 33.71558

Important interactions in final model:

 v1 name1 v2 name2

  2 bio18  1 bio06

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.7107373 

NULL Deviance                : 0.60184

Fit Deviance                 : 0.27889

Explained Deviance           : 0.32295

Percent Deviance Explained   : 53.66

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.15 

Confusion Matrix: 

         observed

predicted   1   0

        1  91  94

        0  10 936

AUC                          :  0.9617 

Percent Correctly Classified :  90.8046 

Sensitivity                  :  0.9009901 

Specificity                  :  0.9087379 

Kappa                        :  0.5888653 

True Skill Statistic         :  0.809728 

   Calibration Statistics 

Intercept (general calibration)                            :  0.3302886 

Slope   (direction and variation in fit)                   :  1.307406 

Testa0b1 (overall reliability of predictors)               :  0.004962452 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9874705 

Testb1|a (refinement given correct calibration)            :  0.001123887 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6093258  (sd 0.12175) 

NULL Deviance                : 0.60179 (sd 0.012753)

Fit Deviance                 : 0.34592 (sd 0.089701)

Explained Deviance           : 0.25586 (sd 0.089989)

Percent Deviance Explained   : 42.511 (sd 14.911)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.143 (sd 0.016364) 

Confusion Matrix: 

         observed

predicted   1   0

        1  83 107

        0  18 923

AUC                          :  0.93221  (sd 0.054087) 

Percent Correctly Classified :  88.94484  (sd 3.8005) 

Sensitivity                  :  0.82  (sd 0.14757) 

Specificity                  :  0.8961817  (sd 0.032286) 

Kappa                        :  0.5191041  (sd 0.14052) 

True Skill Statistic         :  0.7161817  (sd 0.16678) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.3037546  (sd 0.56229) 

Slope   (direction and variation in fit)                   :  1.250301  (sd 0.42313) 

Testa0b1 (overall reliability of predictors)               :  0.4187037  (sd 0.25819) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.6157996  (sd 0.16278) 

Testb1|a (refinement given correct calibration)            :  0.296283  (sd 0.26295) 

Total time = 1.29 min

MARS

Random Forests

MARS Model Results

Data:

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

n(pres)                 =      101

n(abs)                  =      1030

n covariates considered =      4

total time for model fitting = 0.03min

Settings:

random seed used             : 123

mars degree                  : 1

mars penalty                 : 2

Summary of Model:

             nsubsets   gcv    rss

bio18               8 100.0  100.0

bio06               5  53.9   55.9

bio10               4  46.3   48.2

di_all_model        1  20.1   21.3

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6469347 

NULL Deviance                : 0.60184

Fit Deviance                 : 0.3232

Explained Deviance           : 0.27864

Percent Deviance Explained   : 46.298

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.13 

Confusion Matrix: 

         observed

predicted   1   0

        1  88 131

        0  13 899

AUC                          :  0.9369 

Percent Correctly Classified :  87.2679 

Sensitivity                  :  0.8712871 

Specificity                  :  0.8728155 

Kappa                        :  0.4873364 

True Skill Statistic         :  0.7441027 

   Calibration Statistics 

Intercept (general calibration)                            :  0.0003182011 

Slope   (direction and variation in fit)                   :  1.000509 

Testa0b1 (overall reliability of predictors)               :  0.9999836 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9987923 

Testb1|a (refinement given correct calibration)            :  0.9955953 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6251478  (sd 0.11913) 

NULL Deviance                : 0.60179 (sd 0.012753)

Fit Deviance                 : 0.34032 (sd 0.10564)

Explained Deviance           : 0.26146 (sd 0.10463)

Percent Deviance Explained   : 43.475 (sd 17.461)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.124 (sd 0.0069921) 

Confusion Matrix: 

         observed

predicted   1   0

        1  86 137

        0  15 893

AUC                          :  0.93408  (sd 0.048387) 

Percent Correctly Classified :  86.55652  (sd 2.6042) 

Sensitivity                  :  0.85  (sd 0.15811) 

Specificity                  :  0.8670388  (sd 0.026702) 

Kappa                        :  0.4638502  (sd 0.096612) 

True Skill Statistic         :  0.7170388  (sd 0.15586) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.05168208  (sd 0.45429) 

Slope   (direction and variation in fit)                   :  1.208794  (sd 0.70372) 

Testa0b1 (overall reliability of predictors)               :  0.4949496  (sd 0.3601) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5996247  (sd 0.22547) 

Testb1|a (refinement given correct calibration)            :  0.4693118  (sd 0.38895) 

Total time = 0.21 min

Random Forest Modeling Results

Data:

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

n(pres)                 =      101

n(abs)                  =      1030

n covariates considered =      4

total time for model fitting = 0.05min

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

bio18        53.2336 129.8308              88.0748          44.5831

bio06        36.1375  26.3813              46.7594          29.7981

bio10        27.5423  42.1932              45.8815          23.5455

di_all_model 30.6176   1.4534              30.0791          17.5023

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6511067 

NULL Deviance                : 0.60184

Fit Deviance                 : 0.3339

Explained Deviance           : 0.26794

Percent Deviance Explained   : 44.52

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.08 

Confusion Matrix: 

         observed

predicted   1   0

        1  88 139

        0  13 891

AUC                          :  0.9219 

Percent Correctly Classified :  86.56057 

Sensitivity                  :  0.8712871 

Specificity                  :  0.8650485 

Kappa                        :  0.4712255 

True Skill Statistic         :  0.7363357 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.2075747 

Slope   (direction and variation in fit)                   :  0.8276721 

Testa0b1 (overall reliability of predictors)               :  0.05678962 

Testa0|b1(incorrect calibration given correct refinement)  :  0.930109 

Testb1|a (refinement given correct calibration)            :  0.01668593 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6600088  (sd 0.089365) 

NULL Deviance                : 0.60179 (sd 0.012753)

Fit Deviance                 : 0.33047 (sd 0.11212)

Explained Deviance           : 0.27132 (sd 0.11142)

Percent Deviance Explained   : 45.107 (sd 18.567)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.49 (sd 0.01291) 

Confusion Matrix: 

         observed

predicted    1    0

        1   52   21

        0   49 1009

AUC                          :  0.92605  (sd 0.067216) 

Percent Correctly Classified :  93.80356  (sd 1.5287) 

Sensitivity                  :  0.5154545  (sd 0.18847) 

Specificity                  :  0.9796836  (sd 0.0083472) 

Kappa                        :  0.5507925  (sd 0.15201) 

True Skill Statistic         :  0.4951381  (sd 0.18487) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.08164912  (sd 0.38848) 

Slope   (direction and variation in fit)                   :  0.9523096  (sd 0.28146) 

Testa0b1 (overall reliability of predictors)               :  0.5511056  (sd 0.32829) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.5701032  (sd 0.24907) 

Testb1|a (refinement given correct calibration)            :  0.5038549  (sd 0.34535) 

Total time = 0.37 min