Moran, Stuart; He, Yulan and Liu, Kecheng
(2009).
|
Due to copyright restrictions, this file is not available for public download Click here to request a copy from the OU Author. |
| URL: | http://www.iaeng.org/publication/WCE2009/ |
|---|---|
| Google Scholar: | Look up in Google Scholar |
Abstract
Data miners have access to a significant number of classifiers and use them on a variety of different types of dataset. This large selection makes it difficult to know which classifier will perform most effectively in any given case. Usually an understanding of learning algorithms is combined with detailed domain knowledge of the dataset at hand to lead to the choice of a classifier. We propose an empirical framework that quantitatively assesses the accuracy of a selection of classifiers on different datasets, resulting in a set of classification rules generated by the J48 decision tree algorithm. Data miners can follow these rules to select the most effective classifier for their work. By optimising the parameters used for learning and the sampling techniques applied, a set of rules were learned that select with 78% accuracy (with 0.5% classification accuracy tolerance), the most effective classifier.
| Item Type: | Conference Item |
|---|---|
| Copyright Holders: | 2009 International Association of Engineers |
| Extra Information: | Proceedings of the World Congress on Engineering 2009
Editors: S. I. Ao and Len Gelman and David WL Hukins and Andrew Hunter and A. M. Korsunsky ISBN of Vol I (pp1-927): 978-988-17012-5-1 ISBN of Vol II (pp928-1895): 978-988-18210-1-0 Publisher: Newswood Limited Organization: International Association of Engineers |
| Keywords: | Bayesian networks; data mining; classification; search algorithm; decision tree |
| Academic Unit/Department: | Knowledge Media Institute |
| Interdisciplinary Research Centre: | Centre for Research in Computing (CRC) |
| Related URLs: | |
| Item ID: | 25093 |
| Depositing User: | Yulan He |
| Date Deposited: | 07 Dec 2010 09:04 |
| Last Modified: | 26 Oct 2012 04:37 |
| URI: | http://oro.open.ac.uk/id/eprint/25093 |
Actions (login may be required)
| View Item | |
| Public: Report issue / request change |