The performance of the default rule is shown in Table 19 and 20. As shown in Table 19, the default rule is used almost in every different sized training data set. This is roughly the same for the original BruteDL as shown in Table 4. But the number of test observations covered by the default rule is slightly increased on some data sets. For example, the default rule was not used on the 50% and 70% sized training set of iris data by the original BruteDL while it is used by the BruteDL with storing-all-rules. By this comparison, we can say that the default rule usage in the two different versions are not significantly different.
Table 19: Mean Coverage (%) of the Default Rule of BruteDL with
Storing-All-Rules on Different Data Sets
Table 20: Mean Prediction Accuracies (%) of the Default Rule of BruteDL with
Storing-All-Rules on Different Data Sets
