Table of Contents

A.4 Pima Indians Diabetes


A.5 Thyroid Disease

The Thyroid Disease dataset was created by the Garavan Institute, Sydney, Australia. The file was donated by Randolf Werner, obtained from Daimler-Benz. Daimler-Benz obtained the data from J.R. Quinlan. Table 23 shows the test costs for the Thyroid Disease dataset. A nominal cost of $1.00 was assigned to the first 16 tests. The tests in group A share the cost of $2.10 for collecting blood. The FTI test involves a calculation based on the results of the TT4 and T4U tests. This complicates the calculation of the costs of these three tests, so we chose not to use the FTI test in our experiments. The class variable has the values 1 (hypothyroid), 2 (hyperthyroid), and 3 (normal). Table 24 shows the classification costs. There are 3772 cases in this dataset, with no missing values. In our ten random splits, the training sets had 2515 cases and the testing sets had 1257 cases.


Acknowledgments