1. title: Dependent Variable type: Variables

2. Class Variable type: Variable

3. Group Variable type: Variable

4. Method

  options: 
    Custom cut score
    Maximize metric
    Minimize metric
    Maximize metric (LOESS)
    Minimize metric (LOESS)
    Maximize metric (spline)
    Minimize metric (spline)
    Maximize metric (boot)
    Minimize metric (boot)
    Maximize Youden-Index (Kernel smoothed)
    Maximize Youden-Index (Parametric normal)
    default: maximize_metric

Method determines how the optimal cutpoint will be determined after calculating the Metric (below). The default is simply Maximize Metric, which means that the cutpoint with the maximum value for Metric will be chosen as the optimal cutpoint. The options in brackets, LOESS, spline, boot, Kernel smoothed, and Parametric normal refer to different smoothing operations which may be carried out on the metric prior to selecting the respective max/min value.

5. All observed scores

If All observed scores is selected, results table will return statistics for all observed measure scores in the data.

6. DeLong’s Test

If DeLong's Test is selected and two or more groups are provided (either scores on two or more measures or subgroups of a single measure), DeLong’s test for the difference between AUC/ROC will be performed.

7. Specify cut score type: String

Specify cut score will be ignored unless the Custom Cut Score option for Method (above) is chosen. This option is used to return only the results for a single specified observed score. That is, it is only used for filtering results to create tidyier output. As a result of this, the value must be a value in the observed data.

8. Metric type: List

  options:
    Sum Sens-Spec
    Accuracy
    Youden-Index
    Sum: Sens/Spec
    Sum: PPV/NPV
    Prod: Sens/Spec
    Prod: PPV/NPV
    Cohen's Kappa
    Abs. d: Sens/Spec
    ROC
    Abs. d: PPV/NPV
    Chi-squared
    Odds Ratio
    Risk Ratio
    Misclassification Cost
    Total Utility
    F1 score
    

Metric refers to the statistic/value that is to be calculated and used by Method for selecting the optimal cutpoint(s).

9. Bootstrap runs type: Number

Not yet implemented. Please notify the author if this feature would be useful to you

Bootstrap runs is the number of bootstrap samples will be used to assess the variability and the out-of-sample performance. If used, bootstrap samples will be drawn and the optimal cutpoint using method will be determined this way. Additionally, as a way of internal validation, the function in metric will be used to score the out-of-bag predictions using the cutpoints determined by method. Various default metrics are always included in the bootstrap results.

10. Ties

  options:
    All optimal cutpoints
    Mean optimal cutpoint
    Median optimal cutpoint

11. Direction (relative to optimal cutpoint)

  options:
    ">="
    "<="

Direction refers to the direction of the observed score relative to the optimal cutpoint. That is, whether the cutpoint should be a ‘floor’ or a ‘ceiling’.

12. ROC

  default: True
  

Selecting the ROC option will return an ROC curve for each dependent variable and group.

13. Standard error bars

  default: True
  

If ROC has been selected, Standard error bars is used to show approximated standard error ranges using LOESS smoothing.

14. LOESS Smoothing

  default: True

Selecting the LOESS Smoothing option will overlay a LOESS smoothed regression line on the ROC curve.

15. Sensitivity-Specificity Tables

  default: False
  

Sensitivity-Specificity Tables is used to display the raw data at used to calculate sensitivity and specificity. This option may be useful for checking data if the results of the optimal cutpoints or ROC curve are not as expected.