Prototype Generation Based on Instance Filtering and Averaging

Authors:
Chi-Kin Keung;Wai Lam
Affiliations:
-;-
Venue:
PADKK '00 Proceedings of the 4th Pacific-Asia Conference on Knowledge Discovery and Data Mining, Current Issues and New Applications
Year:
2000

Citing 11
Cited 0

Instance-Based Learning Algorithms

Machine Learning
A Nearest Hyperrectangle Learning Method

Machine Learning
Selecting typical instances in instance-based learning

ML92 Proceedings of the ninth international workshop on Machine learning
Tolerating noisy, irrelevant and novel attributes in instance-based learning algorithms

International Journal of Man-Machine Studies - Special issue: symbolic problem solving in noisy and novel task environments
A Weighted Nearest Neighbor Algorithm for Learning with Symbolic Features

Machine Learning
A hybrid nearest-neighbor and nearest-hyperrectangle algorithm

ECML-94 Proceedings of the European conference on machine learning on Machine Learning
Instance Pruning Techniques

ICML '97 Proceedings of the Fourteenth International Conference on Machine Learning
Learning Symbolic Prototypes

ICML '97 Proceedings of the Fourteenth International Conference on Machine Learning
Finding Prototypes For Nearest Neighbor Classifiers

IEEE Transactions on Computers
Symbolic nearest mean classifiers

AAAI'97/IAAI'97 Proceedings of the fourteenth national conference on artificial intelligence and ninth conference on Innovative applications of artificial intelligence
Multiple-prototype classifier design

IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews

Quantified Score

Hi-index	0.00

Visualization

Abstract

We propose a new algorithm, called Prototype Generation and Filtering (PGF), which combines the strength of instance-filtering and instance-averaging techniques. PGF is able to generate representative prototypes while eliminating noise and exceptions. We also introduce a distance measure incorporating the class label entropy information for the prototypes. Experiments have been conducted to compare our PGF algorithm with pure instance filtering, pure instance averaging, as well as state-of-the-art algorithms such as C4.5 and KNN. The results demonstrate that PGF can significantly reduce the size of the data while maintaining and even improving the classification performance.