A method for optimal division of data sets for use in neural networks

Authors:
Patricia S. Crowther;Robert J. Cox
Affiliations:
School of Information Sciences and Engineering, University of Canberra, Australia;School of Information Sciences and Engineering, University of Canberra, Australia
Venue:
KES'05 Proceedings of the 9th international conference on Knowledge-Based Intelligent Information and Engineering Systems - Volume Part IV
Year:
2005

Citing 3
Cited 6

Bagging predictors

Machine Learning
Neural Networks: A Comprehensive Foundation

Neural Networks: A Comprehensive Foundation
An Investigation into the Effect of Ensemble Size and Voting Threshold on the Accuracy of Neural Network Ensembles

AI '99 Proceedings of the 12th Australian Joint Conference on Artificial Intelligence: Advanced Topics in Artificial Intelligence

Symmetry enhanced adaboost

ISVC'10 Proceedings of the 6th international conference on Advances in visual computing - Volume Part I
Development of a parsimonious GA-NN ensemble model with a case study for Charpy impact energy prediction

Advances in Engineering Software
PCA enhanced training data for adaboost

CAIP'11 Proceedings of the 14th international conference on Computer analysis of images and patterns - Volume Part I
Implicit scene context for object segmentation and classification

DAGM'11 Proceedings of the 33rd international conference on Pattern recognition
Accuracy of neural network classifiers as a property of the size of the data set

KES'06 Proceedings of the 10th international conference on Knowledge-Based Intelligent Information and Engineering Systems - Volume Part III
Segmentation and classification of objects with implicit scene context

Proceedings of the 15th international conference on Theoretical Foundations of Computer Vision: outdoor and large-scale real-world scene analysis

Quantified Score

Hi-index	0.00

Visualization

Abstract

Neural Networks are used to find a generalised solution from a sample set of a problem domain. When a small sample is all that is available, the correct division of data between the training, testing and validation sets is crucial to the performance of the resultant trained network. Data is often divided uniformly between the three data sets. We propose an alternative method for the optimal division of the data, based on empirical evidence from experiments with artificial data. The method is tested on real world data sets, with encouraging results.