Skip to content

MedPlot

To create general graph please use createHtmlGraph function in MedPlot. to create ROC graphs use plotAUC function.   plotAUC** Input: **

  • vector of all models predictions vector(each record is prediction vector result of specific model)
  • vector of the labels
  • vector of the models\predictors names (same size as the first vector)
  • output direcotry for writing graphs

  • optional - indexes is a mask for selecting rows and may be omitted to select all   plotAUC Output:** ** the function outputs the following graphs:

  • ROC graph curve

  • PPV graph curve - for each false positive rate - it's PPV value for each model
  • False Positive rate as function of model score - you may see each model scores VS it's false positive rate
  • Label distribution of cases and controls  For Each Graph you may search for specific working point in the search button - fill in X value (false positive rate) and it will find you the Y value (Sensitivity or PPV depends on the graph)   Example Run:   
    1
2
3
4
    vector<float> preds; //from MedModel scores
vector<float> age_baseline; //the age_baseline for each sample
vector<float> labels;
plotAUC({age_baseline, preds}, labels, {"Age_Baseline" ,"MedModel"}, "~/Auc_Example_Folder")