Species 102: Pinus aristata

General Linear Models

Boosted Regression Trees

Generalized Linear Model Results

Data:

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

n(pres)                 =      129

n(abs)                  =      3207

n covariates considered =      5

total time for model fitting = 0.33min

Settings:

model family          : binomial

simplification method : AIC

Results:

number covariates in final model   : 10

Call:

glm(formula = response ~ I(bio05^2) + bio05 + bio15 + I(bio15^2) + 

    bio04 + I(bio04^2) + I(di_all_model^2) + di_all_model + bio05:bio15 + 

    bio04:di_all_model, family = model.family, data = dat, weights = weight, 

    na.action = "na.exclude")

Deviance Residuals: 

     Min        1Q    Median        3Q       Max  

-2.54384  -0.03654  -0.00023   0.00000   2.77488  

Coefficients:

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

(Intercept)        -1.637e+02  2.881e+01  -5.684 1.32e-08 ***

I(bio05^2)         -1.122e-03  3.289e-04  -3.412 0.000644 ***

bio05               4.233e-01  1.082e-01   3.912 9.16e-05 ***

bio15               1.030e+00  2.250e-01   4.576 4.74e-06 ***

I(bio15^2)         -6.799e-03  1.582e-03  -4.298 1.72e-05 ***

bio04               2.787e-02  7.681e-03   3.628 0.000285 ***

I(bio04^2)         -1.733e-06  5.106e-07  -3.395 0.000686 ***

I(di_all_model^2)  -2.761e-03  1.083e-03  -2.550 0.010777 *  

di_all_model       -1.609e-01  1.766e-01  -0.911 0.362248    

bio05:bio15        -1.938e-03  1.326e-03  -1.461 0.143894    

bio04:di_all_model  3.504e-05  2.520e-05   1.390 0.164451    

---

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

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 1092.15  on 3335  degrees of freedom

Residual deviance:  438.18  on 3325  degrees of freedom

AIC: 460.18

Number of Fisher Scoring iterations: 12

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6809785 

NULL Deviance                : 0.32738

Fit Deviance                 : 0.13135

Explained Deviance           : 0.19603

Percent Deviance Explained   : 59.879

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.08 

Confusion Matrix: 

         observed

predicted    1    0

        1  120  221

        0    9 2986

AUC                          :  0.9802 

Percent Correctly Classified :  93.10552 

Sensitivity                  :  0.9302326 

Specificity                  :  0.9310882 

Kappa                        :  0.4815472 

True Skill Statistic         :  0.8613208 

   Calibration Statistics 

Intercept (general calibration)                            :  0.001391534 

Slope   (direction and variation in fit)                   :  1.001609 

Testa0b1 (overall reliability of predictors)               :  0.9997936 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9967012 

Testb1|a (refinement given correct calibration)            :  0.9841301 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6639302  (sd 0.069466) 

NULL Deviance                : 0.32736 (sd 0.0060144)

Fit Deviance                 : 0.14541 (sd 0.02409)

Explained Deviance           : 0.18195 (sd 0.025384)

Percent Deviance Explained   : 55.554 (sd 7.4641)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.078 (sd 0.0042164) 

Confusion Matrix: 

         observed

predicted    1    0

        1  116  242

        0   13 2965

AUC                          :  0.97404  (sd 0.010574) 

Percent Correctly Classified :  92.35529  (sd 2.2145) 

Sensitivity                  :  0.8987179  (sd 0.07413) 

Specificity                  :  0.9245308  (sd 0.022354) 

Kappa                        :  0.4547697  (sd 0.089189) 

True Skill Statistic         :  0.8232487  (sd 0.080835) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.03719519  (sd 0.72822) 

Slope   (direction and variation in fit)                   :  1.013631  (sd 0.37143) 

Testa0b1 (overall reliability of predictors)               :  0.3727603  (sd 0.30393) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.4981355  (sd 0.31578) 

Testb1|a (refinement given correct calibration)            :  0.3730645  (sd 0.3133) 

Total time = 3.19 min

Boosted Regression Tree Modeling Results

Data:

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

n(pres)                 =      129

n(abs)                  =      3207

n covariates considered =      5

total time for model fitting = 0.39min

Settings:

random seed used             : 19680828

tree complexity              : 7

learning rate                : 0.0514

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

 bio15 51.7266

 bio05 48.2734

Important interactions in final model:

 v1 name1 v2 name2

  2 bio15  1 bio05

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.7992958 

NULL Deviance                : 0.32738

Fit Deviance                 : 0.092788

Explained Deviance           : 0.23459

Percent Deviance Explained   : 71.658

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.13 

Confusion Matrix: 

         observed

predicted    1    0

        1  124  126

        0    5 3081

AUC                          :  0.9918 

Percent Correctly Classified :  96.07314 

Sensitivity                  :  0.9612403 

Specificity                  :  0.9607109 

Kappa                        :  0.6357727 

True Skill Statistic         :  0.9219513 

   Calibration Statistics 

Intercept (general calibration)                            :  0.5867236 

Slope   (direction and variation in fit)                   :  1.629895 

Testa0b1 (overall reliability of predictors)               :  8.08914e-07 

Testa0|b1(incorrect calibration given correct refinement)  :  0.8709776 

Testb1|a (refinement given correct calibration)            :  1.195254e-07 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.5529046  (sd 0.071792) 

NULL Deviance                : 0.32736 (sd 0.0060144)

Fit Deviance                 : 0.18154 (sd 0.021979)

Explained Deviance           : 0.14582 (sd 0.02046)

Percent Deviance Explained   : 44.572 (sd 6.3726)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.149 (sd 0.071872) 

Confusion Matrix: 

         observed

predicted    1    0

        1   93  155

        0   36 3052

AUC                          :  0.95957  (sd 0.017663) 

Percent Correctly Classified :  94.27433  (sd 1.4742) 

Sensitivity                  :  0.7211538  (sd 0.0894) 

Specificity                  :  0.9516706  (sd 0.015856) 

Kappa                        :  0.4736283  (sd 0.085419) 

True Skill Statistic         :  0.6728244  (sd 0.087012) 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.3060953  (sd 0.37264) 

Slope   (direction and variation in fit)                   :  0.7358548  (sd 0.12951) 

Testa0b1 (overall reliability of predictors)               :  0.218151  (sd 0.23905) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.4350402  (sd 0.19389) 

Testb1|a (refinement given correct calibration)            :  0.1833048  (sd 0.26647) 

Total time = 4.12 min

MARS

Random Forests

MARS Model Results

Data:

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

n(pres)                 =      129

n(abs)                  =      3207

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

bio05              13 100.0  100.0

bio15              12  77.6   78.6

bio02              10  44.2   47.3

bio04               9  36.4   40.1

di_all_model        3  14.0   17.3

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.6762435 

NULL Deviance                : 0.32738

Fit Deviance                 : 0.13263

Explained Deviance           : 0.19475

Percent Deviance Explained   : 59.487

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.08 

Confusion Matrix: 

         observed

predicted    1    0

        1  118  241

        0   11 2966

AUC                          :  0.9795 

Percent Correctly Classified :  92.44604 

Sensitivity                  :  0.9147287 

Specificity                  :  0.9248519 

Kappa                        :  0.4524544 

True Skill Statistic         :  0.8395806 

   Calibration Statistics 

Intercept (general calibration)                            :  0.001729788 

Slope   (direction and variation in fit)                   :  1.00195 

Testa0b1 (overall reliability of predictors)               :  0.9997294 

Testa0|b1(incorrect calibration given correct refinement)  :  0.9967392 

Testb1|a (refinement given correct calibration)            :  0.9817259 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6661717  (sd 0.071773) 

NULL Deviance                : 0.32736 (sd 0.0060144)

Fit Deviance                 : 0.14109 (sd 0.023095)

Explained Deviance           : 0.18627 (sd 0.022529)

Percent Deviance Explained   : 56.914 (sd 6.8873)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.084 (sd 0.005164) 

Confusion Matrix: 

         observed

predicted    1    0

        1  118  240

        0   11 2967

AUC                          :  0.97596  (sd 0.0099093) 

Percent Correctly Classified :  92.47496  (sd 1.6128) 

Sensitivity                  :  0.9147436  (sd 0.07645) 

Specificity                  :  0.9251509  (sd 0.016104) 

Kappa                        :  0.458368  (sd 0.070992) 

True Skill Statistic         :  0.8398945  (sd 0.080109) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.09634112  (sd 0.75317) 

Slope   (direction and variation in fit)                   :  1.079183  (sd 0.43074) 

Testa0b1 (overall reliability of predictors)               :  0.4580036  (sd 0.32203) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.54182  (sd 0.31212) 

Testb1|a (refinement given correct calibration)            :  0.4009488  (sd 0.31124) 

Total time = 0.93 min

Random Forest Modeling Results

Data:

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

n(pres)                 =      129

n(abs)                  =      3207

n covariates considered =      5

total time for model fitting = 0.16min

Settings:

random seed used                       : 19680828

n covariates considered at each split  : 5

n trees                                : 1000

Relative performance of predictors in final model:

                    0        1 MeanDecreaseAccuracy MeanDecreaseGini

bio15        118.5182  23.2735             137.6064          46.5308

bio04        102.0131  10.0829             105.7808          34.8322

bio05         79.0261 166.3109              98.7315          41.2559

bio02         10.7206  36.7563              34.6169          19.4253

di_all_model  29.4343  -8.2210              22.6990          14.2568

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

Evaluation Statistics applied to train split:

Correlation Coefficient      : 0.674822 

NULL Deviance                : 0.32738

Fit Deviance                 : 0.14412

Explained Deviance           : 0.18327

Percent Deviance Explained   : 55.979

  Threshold Methods based on Sens=Spec 

Threshold                    :  0.04 

Confusion Matrix: 

         observed

predicted    1    0

        1  120  259

        0    9 2948

AUC                          :  0.9737 

Percent Correctly Classified :  91.96643 

Sensitivity                  :  0.9302326 

Specificity                  :  0.9192392 

Kappa                        :  0.4401374 

True Skill Statistic         :  0.8494717 

   Calibration Statistics 

Intercept (general calibration)                            :  -0.1818857 

Slope   (direction and variation in fit)                   :  0.9510659 

Testa0b1 (overall reliability of predictors)               :  0.4289002 

Testa0|b1(incorrect calibration given correct refinement)  :  0.3290181 

Testb1|a (refinement given correct calibration)            :  0.3895638 

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

Evaluation Statistics applied to crossValidation split:

Correlation Coefficient      : 0.6864972  (sd 0.057226) 

NULL Deviance                : 0.32736 (sd 0.0060144)

Fit Deviance                 : 0.13593 (sd 0.02149)

Explained Deviance           : 0.19142 (sd 0.022522)

Percent Deviance Explained   : 58.458 (sd 6.6089)

  Threshold Methods based on Sens=Spec 

Mean Threshold               : 0.478 (sd 0.011353) 

Confusion Matrix: 

         observed

predicted    1    0

        1   66   26

        0   63 3181

AUC                          :  0.9749  (sd 0.018239) 

Percent Correctly Classified :  97.33227  (sd 0.53627) 

Sensitivity                  :  0.5121795  (sd 0.099878) 

Specificity                  :  0.9918915  (sd 0.0039453) 

Kappa                        :  0.5818576  (sd 0.087561) 

True Skill Statistic         :  0.504071  (sd 0.099783) 

   Calibration Statistics 

Intercept (general calibration)                            :  0.1974522  (sd 0.49581) 

Slope   (direction and variation in fit)                   :  1.158184  (sd 0.27426) 

Testa0b1 (overall reliability of predictors)               :  0.6099781  (sd 0.30734) 

Testa0|b1(incorrect calibration given correct refinement)  :  0.6688844  (sd 0.26016) 

Testb1|a (refinement given correct calibration)            :  0.5061906  (sd 0.32832) 

Total time = 1.85 min