Table of Contents

4.2.1 All Tests Immediate


4.2.2 No Group Discounts

Another hypothesis is that ICET is superior simply because it can handle groups of tests that share a common cost. In this experiment, we eliminated group discounts for tests that share a common cost. That is, test costs were not conditional on prior tests. Otherwise, nothing changed from the baseline experiments. Table 8 summarizes the results of the experiment. ICET maintains its advantage over the other algorithms.


4.2.3 Complex Classification Cost Matrices