Clustering data manipulation methods for the development of local specialists

Authors:
Matthew J. Spencer;Tim Whitfort;John McCullagh
Affiliations:
Department of Computer Science and Computer Engineering, La Trobe University, Bendigo, VIC, Australia;Department of Computer Science and Computer Engineering, La Trobe University, Bendigo, VIC, Australia;Department of Computer Science and Computer Engineering, La Trobe University, Bendigo, VIC, Australia
Venue:
AIAP'07 Proceedings of the 25th conference on Proceedings of the 25th IASTED International Multi-Conference: artificial intelligence and applications
Year:
2007

Citing 7
Cited 0

Bagging predictors

Machine Learning
Analysis of Linear and Order Statistics Combiners for Fusion of Imbalanced Classifiers

MCS '02 Proceedings of the Third International Workshop on Multiple Classifier Systems
Dynamic ensemble approach for estimating organic carbon using computational intelligence

ACST'06 Proceedings of the 2nd IASTED international conference on Advances in computer science and technology
Boosting with averaged weight vectors

MCS'03 Proceedings of the 4th international conference on Multiple classifier systems
Clustering data manipulation method for ensembles

AI'06 Proceedings of the 19th Australian joint conference on Artificial Intelligence: advances in Artificial Intelligence
Switching between selection and fusion in combining classifiers: anexperiment

IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
A constructive algorithm for training cooperative neural network ensembles

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

The success of an ensemble critically depends on the development of its members. Ensemble members need to be both diverse and accurate to improve in performance compared to a single member. A number of data manipulation techniques have been used to develop diverse members, such as bagging and boosting. These methods have demonstrated that distibuting different training patterns to each member and aggregating the estimates can improve performance. Clustering techniques have shown potential in developing local specialists to assist in difficult regions where a generalist performs poorly. This paper presents two clustering methods for the local selection of training data. The results demonstrate that a combination of local specialists can lead to an improvement in ensemble performance.