Chapter 10 Decision trees and Markov chains for gene finding

Abstract

Decision trees are an important and a widely used technique for classifying data. Recently, they have been adapted for use in computational biology, especially for the analysis of DNA sequence data. This chapter describes how decision trees are used in the gene-finding system and how they might be used for other problems of sequence analysis. A decision tree is a collection of nodes and edges, where each node tests some feature or features of the data. These trees are used as classifiers by passing an example down from the root node, which is a specially designated node in the tree. An example is tested at each node; it proceeds down the left or right branch depending on the outcome of the test. The chapter explains the basic tree-induction algorithm; the result of this algorithm is a decision tree with test nodes and leaf nodes. The leaf nodes contain class labels.