A new method for discretization of continuous attributes based on VPRS

Authors:
Jin-Mao Wei;Guo-Ying Wang;Xiang-Ming Kong;Shu-Jie Li;Shu-Qin Wang;Da-You Liu
Affiliations:
Institute of Computational Intelligence, Northeast Normal University, Changchun, Jilin, China;Institute of Computational Intelligence, Northeast Normal University, Changchun, Jilin, China;Institute of Computational Intelligence, Northeast Normal University, Changchun, Jilin, China;Institute of Computational Intelligence, Northeast Normal University, Changchun, Jilin, China;School of Mathematics & Statistics, Northeast Normal University, Changchun, Jilin, China;Open Symbol Computation and Knowledge Engineering Laboratory of State Education, Jilin University, Changchun, Jilin, China
Venue:
RSCTC'06 Proceedings of the 5th international conference on Rough Sets and Current Trends in Computing
Year:
2006

Citing 8
Cited 1

Simplifying decision trees

International Journal of Man-Machine Studies - Special Issue: Knowledge Acquisition for Knowledge-based Systems. Part 5
Variable precision rough set model

Journal of Computer and System Sciences
Rough Sets: Theoretical Aspects of Reasoning about Data

Rough Sets: Theoretical Aspects of Reasoning about Data
Discretization: An Enabling Technique

Data Mining and Knowledge Discovery
Use of Contextual Information for Feature Ranking and Discretization

IEEE Transactions on Knowledge and Data Engineering
Chi2: Feature Selection and Discretization of Numeric Attributes

TAI '95 Proceedings of the Seventh International Conference on Tools with Artificial Intelligence
On Efficient Handling of Continuous Attributes in Large Data Bases

Fundamenta Informaticae
Improved use of continuous attributes in C4.5

Journal of Artificial Intelligence Research

1-vs-others rough decision forest

RSKT'10 Proceedings of the 5th international conference on Rough set and knowledge technology

Quantified Score

Hi-index	0.00

Visualization

Abstract

A new method for discretization of continuous features based on the Variable Precision Rough Set theory is proposed and tested in the process of inducing decision trees. Through rectifying error ratio, the generalization capability of decision trees is enhanced by enlarging or reducing the sizes of positive regions. Two ways of computing frequency and width are deployed to calculate the misclassifying rate of the data, and thus the negative effect on decision trees is reduced, by which the discretization points are determined. In the paper, we use some open data sets to testify the method. The results are compared with that obtained by C4.5, which shows that the presented method is a feasible way to discretization of continuous features in applications.