Investigating the performance of MLP classifiers where limited training data are available for some classes

CR Parikh; MJ Pont; Yuhua Li; NB Jones

Investigating the performance of MLP classifiers where limited training data are available for some classes

CR Parikh, MJ Pont, Yuhua Li, NB Jones

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The standard implementation of the back-propagation training algorithm for multi-layer Perceptron (MLP) neural networks assumes that there are equal number of samples for training each of the required classes. Where limited training data are available for one (or more) classes, sub-optimal performance may be obtained. We have demonstrated in a previous study [Parikh ct al., 1999. Proceedings of Condition Monitoring 1999, Swansea, UK] that, where unequal training class cannot be avoided, performance of the classifier may be substantially improved by duplicating the available patterns in the smaller class. In this study, we investigate whether the addition of random noise to the `duplicated' training patterns will further improve the classification performance. In the study conducted here, we conclude that the addition of noise does not give a consistent improvement in performance.

Language	English
Title of host publication	Unknown Host Publication
Pages	22-27
Number of pages	6
Publication status	Published - 2000
Event	SOFT COMPUTING TECHNIQUES AND APPLICATIONS - Duration: 1 Jan 2000 → …

Publication series

Name	ADVANCES IN SOFT COMPUTING

Conference

Conference	SOFT COMPUTING TECHNIQUES AND APPLICATIONS
Period	1/01/00 → …

Fingerprint

Condition monitoring

Multilayer neural networks

Backpropagation

Classifiers

Neural networks

Keywords

Multi-layer Perceptron
training algorithm
condition monitoring
fault diagnosis

Cite this

Parikh, CR., Pont, MJ., Li, Y., & Jones, NB. (2000). Investigating the performance of MLP classifiers where limited training data are available for some classes. In Unknown Host Publication (pp. 22-27). (ADVANCES IN SOFT COMPUTING).

Parikh, CR ; Pont, MJ ; Li, Yuhua ; Jones, NB. / Investigating the performance of MLP classifiers where limited training data are available for some classes. Unknown Host Publication. 2000. pp. 22-27 (ADVANCES IN SOFT COMPUTING).

@inproceedings{951dceb86ad74d869015e62322cfb539,

title = "Investigating the performance of MLP classifiers where limited training data are available for some classes",

abstract = "The standard implementation of the back-propagation training algorithm for multi-layer Perceptron (MLP) neural networks assumes that there are equal number of samples for training each of the required classes. Where limited training data are available for one (or more) classes, sub-optimal performance may be obtained. We have demonstrated in a previous study [Parikh ct al., 1999. Proceedings of Condition Monitoring 1999, Swansea, UK] that, where unequal training class cannot be avoided, performance of the classifier may be substantially improved by duplicating the available patterns in the smaller class. In this study, we investigate whether the addition of random noise to the `duplicated' training patterns will further improve the classification performance. In the study conducted here, we conclude that the addition of noise does not give a consistent improvement in performance.",

keywords = "Multi-layer Perceptron, training algorithm, condition monitoring, fault diagnosis",

author = "CR Parikh and MJ Pont and Yuhua Li and NB Jones",

note = "Workshop 99 on Recent Advances in Soft Computing, LEICESTER, ENGLAND, JUL 01-02, 1999",

year = "2000",

language = "English",

series = "ADVANCES IN SOFT COMPUTING",

pages = "22--27",

booktitle = "Unknown Host Publication",

}

Parikh, CR, Pont, MJ, Li, Y & Jones, NB 2000, Investigating the performance of MLP classifiers where limited training data are available for some classes. in Unknown Host Publication. ADVANCES IN SOFT COMPUTING, pp. 22-27, SOFT COMPUTING TECHNIQUES AND APPLICATIONS, 1/01/00.

Investigating the performance of MLP classifiers where limited training data are available for some classes. / Parikh, CR; Pont, MJ; Li, Yuhua; Jones, NB.

Unknown Host Publication. 2000. p. 22-27 (ADVANCES IN SOFT COMPUTING).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Investigating the performance of MLP classifiers where limited training data are available for some classes

AU - Parikh, CR

AU - Pont, MJ

AU - Li, Yuhua

AU - Jones, NB

N1 - Workshop 99 on Recent Advances in Soft Computing, LEICESTER, ENGLAND, JUL 01-02, 1999

PY - 2000

Y1 - 2000

N2 - The standard implementation of the back-propagation training algorithm for multi-layer Perceptron (MLP) neural networks assumes that there are equal number of samples for training each of the required classes. Where limited training data are available for one (or more) classes, sub-optimal performance may be obtained. We have demonstrated in a previous study [Parikh ct al., 1999. Proceedings of Condition Monitoring 1999, Swansea, UK] that, where unequal training class cannot be avoided, performance of the classifier may be substantially improved by duplicating the available patterns in the smaller class. In this study, we investigate whether the addition of random noise to the `duplicated' training patterns will further improve the classification performance. In the study conducted here, we conclude that the addition of noise does not give a consistent improvement in performance.

AB - The standard implementation of the back-propagation training algorithm for multi-layer Perceptron (MLP) neural networks assumes that there are equal number of samples for training each of the required classes. Where limited training data are available for one (or more) classes, sub-optimal performance may be obtained. We have demonstrated in a previous study [Parikh ct al., 1999. Proceedings of Condition Monitoring 1999, Swansea, UK] that, where unequal training class cannot be avoided, performance of the classifier may be substantially improved by duplicating the available patterns in the smaller class. In this study, we investigate whether the addition of random noise to the `duplicated' training patterns will further improve the classification performance. In the study conducted here, we conclude that the addition of noise does not give a consistent improvement in performance.

KW - Multi-layer Perceptron

KW - training algorithm

KW - condition monitoring

KW - fault diagnosis

M3 - Conference contribution

T3 - ADVANCES IN SOFT COMPUTING

SP - 22

EP - 27

BT - Unknown Host Publication

ER -

Parikh CR, Pont MJ, Li Y, Jones NB. Investigating the performance of MLP classifiers where limited training data are available for some classes. In Unknown Host Publication. 2000. p. 22-27. (ADVANCES IN SOFT COMPUTING).