Now that we know the important variables, it is time to examine how the OneR package outputs the decision rules. Since glucose was deemed the single best predictor, the decision rule for predicting diabetes is shown under the Rules: section, and the optimal cutoff point (139) for glucose is displayed in the contingency table as shown below.
Now, look at the confusion matrix contingency tables-one for the absolute counts and the other for the relative percentages.
For our example, a glucose level greater or equal to 139 predicts diabetes in 18% of the total cases. However, note that most of the predictive power comes from ...
