Using Grouping and Uncertain Reasoning during ID3 Decision Tree Construction and Testing

Abstract

Quinlan's ID3 algorithm produces classification trees that perform poorly when used with noisy or continuous-valued data. The UR-ID3 algorithm combines uncertain reasoning with the IDS rule set to handle approximate test data. This paper presents a variation, called UR-ID3e, that assumes that training data is also uncertain. Experimental results are presented that compare the performance of UR-ID3e with other algorithms.

Department(s)

Mathematics and Statistics

Document Type

Article - Journal

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2024 The Authors, All rights reserved.

Publication Date

01 Dec 1996

This document is currently not available here.

Share

 
COinS