Random Forest

Model with Five RTL Group Levels

Metrics before Tuning

The initial model run generates an roc_auc of 0.586. Before tuning, this model is already stronger than the decision tree.

## # A tibble: 1 × 6
##   .metric .estimator  mean     n std_err .config             
##   <chr>   <chr>      <dbl> <int>   <dbl> <chr>               
## 1 roc_auc hand_till  0.586    10  0.0138 Preprocessor1_Model1

Metrics after Tuning

With tuning, the roc_auc value improves to 0.599

## # A tibble: 5 × 8
##    mtry min_n .metric .estimator  mean     n std_err .config              
##   <int> <int> <chr>   <chr>      <dbl> <int>   <dbl> <chr>                
## 1     4    24 roc_auc hand_till  0.599    10  0.0111 Preprocessor1_Model03
## 2     4    25 roc_auc hand_till  0.596    10  0.0129 Preprocessor1_Model08
## 3     3    30 roc_auc hand_till  0.594    10  0.0120 Preprocessor1_Model01
## 4     3    17 roc_auc hand_till  0.594    10  0.0129 Preprocessor1_Model09
## 5     2     3 roc_auc hand_till  0.590    10  0.0161 Preprocessor1_Model04

Model with Workflow Added

With the workflow, the roc_auc eclipses 0.601, the best model result yet.

## [[1]]
## # A tibble: 2 × 4
##   .metric  .estimator .estimate .config             
##   <chr>    <chr>          <dbl> <chr>               
## 1 accuracy multiclass     0.366 Preprocessor1_Model1
## 2 roc_auc  hand_till      0.601 Preprocessor1_Model1

Important Predictor Plot

Predictions

## # A tibble: 191 × 1
##    .pred_class
##    <fct>      
##  1 7-12       
##  2 13-18      
##  3 7-12       
##  4 7-12       
##  5 7-12       
##  6 7-12       
##  7 7-12       
##  8 7-12       
##  9 13-18      
## 10 7-12       
## # … with 181 more rows

Model with Two RTL Group Levels

Metrics before Tuning

The initial model run generates an roc_auc of 0.555. Not as strong as the 5-level model.

## # A tibble: 1 × 6
##   .metric .estimator  mean     n std_err .config             
##   <chr>   <chr>      <dbl> <int>   <dbl> <chr>               
## 1 roc_auc binary     0.555    10  0.0202 Preprocessor1_Model1

Metrics after Tuning

With tuning, the roc_auc value improves to 0.557

## # A tibble: 5 × 8
##    mtry min_n .metric .estimator  mean     n std_err .config              
##   <int> <int> <chr>   <chr>      <dbl> <int>   <dbl> <chr>                
## 1     2    13 roc_auc binary     0.557    10  0.0198 Preprocessor1_Model06
## 2     4    25 roc_auc binary     0.556    10  0.0204 Preprocessor1_Model08
## 3     4    24 roc_auc binary     0.556    10  0.0203 Preprocessor1_Model03
## 4     3    17 roc_auc binary     0.554    10  0.0195 Preprocessor1_Model09
## 5     2     3 roc_auc binary     0.550    10  0.0188 Preprocessor1_Model04

Model with Workflow Added

With the workflow, the roc_auc generates a value of 0.554

## [[1]]
## # A tibble: 2 × 4
##   .metric  .estimator .estimate .config             
##   <chr>    <chr>          <dbl> <chr>               
## 1 accuracy binary         0.605 Preprocessor1_Model1
## 2 roc_auc  binary         0.554 Preprocessor1_Model1

Important Predictor Plot

Predictions

## # A tibble: 190 × 1
##    .pred_class
##    <fct>      
##  1 0-12       
##  2 0-12       
##  3 13-30      
##  4 0-12       
##  5 0-12       
##  6 0-12       
##  7 0-12       
##  8 0-12       
##  9 0-12       
## 10 0-12       
## # … with 180 more rows