2003Unpublished venueRequires access

Research on algorithm of decision tree induction

Hua Ding, Xiu-Kun Wang

Open publisher page 8 citations

Abstract

This paper introduces the development trend of decision trees in the past several years. During these years, the ID3 algorithm has been at its highest point in decision trees. ID3 makes use of information entropy as heuristics to select a "excellent attribution" at each node, so it can get a smaller depth but not the proper width of the tree, i.e., if the width is large, no matter how small the decision tree is, the leaf node will not be small. It is very important that smaller nodes get higher classify precision for decision trees. This paper provides two algorithms, which can avoid the deficiency of ID3 and reduce the width of the tree to get a better result - the one is probability based PID, and the other is entropy based EMID.

About this research paper

What this paper is about

This paper introduces the development trend of decision trees in the past several years. During these years, the ID3 algorithm has been at its highest point in decision trees. ID3 makes use of information entropy as heuristics to select a "excellent attribution" at each node, so it can get a smaller depth but not the proper width of the tree, i.e., if the width is large, no matter how small the decision tree is, the leaf node will not be small. It is very important that smaller nodes get higher classify precision for decision trees. This paper provides two algorithms, which can avoid the deficiency of ID3 and reduce the width of the tree to get a better result - the one is probability based PID, and the other is entropy based EMID.

Why it matters

OpenAlex reports 8 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

This paper introduces the development trend of decision trees in the past several years. During these years, the ID3 algorithm has been at its highest point in decision trees. ID3 makes use of information entropy as heuristics to select a "excellent attribution" at each node, so it can get a smaller depth but not the proper width of the tree, i.e., if the width is large, no matter how small the decision tree is, the leaf node will not be small. It is very important that smaller nodes get higher classify precision for decision trees. This paper provides two algorithms, which can avoid the deficiency of ID3 and reduce the width of the tree to get a better result - the one is probability based PID, and the other is entropy based EMID.

Key concepts: ID3 algorithm, Decision tree, Incremental decision tree, Heuristics, Decision tree learning, ID3, Computer science, Decision stump

Related papers

Back to paper searchBrowse research topicsOriginal source
Research on algorithm of decision tree induction — Research Paper | ScholarLens