ABSTRACT
CONTENTS 29.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487 29.2 A Secondary Node-Splitting Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488 29.3 The Formation of a Deterministic Forest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488 29.4 Comparison Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489 29.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
1To whom reprint requests should be addressed. E-mail: heping.zhang@yale.edu
29.1 Introduction Microarray technology has unveiled a great opportunity for tumor and cancer classifications [7, 5, 10, 12, 17]. The classic approach does not discriminate among tumors with similar histopathologic features, which may vary in clinical course and in response to treatment [7]. Microarray data monitor gene expression profiles from thousands of genes simultaneously. Appropriate use of such rich information can lead to improvement in the classification and diagnosis of cancer over the classic morphologic approaches.
