A high performance k-NN approach using binary neural networks

V J Hodge; K J Lees; J L Austin

doi:10.1016/j.neunet.2003.11.008

A high performance k-NN approach using binary neural networks

V J Hodge, K J Lees, J L Austin

Computer Science

Research output: Contribution to journal › Article › peer-review

Abstract

This paper evaluates a novel k-nearest neighbour (k-NN) classifier built from binary neural networks. The binary neural approach uses robust encoding to map standard ordinal, categorical and numeric data sets onto a binary neural network. The binary neural network uses high speed pattern matching to recall a candidate set of matching records, which are then processed by a conventional k-NN approach to determine the k-best matches. We compare various configurations of the binary approach to a conventional approach for memory overheads, training speed, retrieval speed and retrieval accuracy. We demonstrate the superior performance with respect to speed and memory requirements of the binary approach compared to the standard approach and we pinpoint the optimal configurations. (C) 2003 Elsevier Ltd. All rights reserved.

Original language	English
Pages (from-to)	441-458
Number of pages	17
Journal	Neural Networks
Volume	17
Issue number	3
DOIs	https://doi.org/10.1016/j.neunet.2003.11.008
Publication status	Published - Apr 2004

Bibliographical note

Copyright © 2003 Elsevier Ltd. This is an author produced version of a paper published in Neural Networks. This paper has been peer-reviewed but does not include the final publisher proof-corrections or journal pagination.

Keywords

binary neural network
associative memory
correlation matrix memory
k-nearest neighbour
euclidean distance
robust encoding
quantisation
binary mapping

Access to Document

10.1016/j.neunet.2003.11.008

hodgevj5.pdf

http://dx.doi.org/10.1016/j.neunet.2003.11.008

Cite this

@article{96a64c45c0e84d5ebc4119aa14a60fae,

title = "A high performance k-NN approach using binary neural networks",

abstract = "This paper evaluates a novel k-nearest neighbour (k-NN) classifier built from binary neural networks. The binary neural approach uses robust encoding to map standard ordinal, categorical and numeric data sets onto a binary neural network. The binary neural network uses high speed pattern matching to recall a candidate set of matching records, which are then processed by a conventional k-NN approach to determine the k-best matches. We compare various configurations of the binary approach to a conventional approach for memory overheads, training speed, retrieval speed and retrieval accuracy. We demonstrate the superior performance with respect to speed and memory requirements of the binary approach compared to the standard approach and we pinpoint the optimal configurations. (C) 2003 Elsevier Ltd. All rights reserved.",

keywords = "binary neural network, associative memory, correlation matrix memory, k-nearest neighbour, euclidean distance, robust encoding, quantisation, binary mapping",

author = "Hodge, {V J} and Lees, {K J} and Austin, {J L}",

note = "Copyright {\textcopyright} 2003 Elsevier Ltd. This is an author produced version of a paper published in Neural Networks. This paper has been peer-reviewed but does not include the final publisher proof-corrections or journal pagination.",

year = "2004",

month = apr,

doi = "10.1016/j.neunet.2003.11.008",

language = "English",

volume = "17",

pages = "441--458",

journal = "Neural Networks",

issn = "0893-6080",

publisher = "Elsevier",

number = "3",

}

TY - JOUR

T1 - A high performance k-NN approach using binary neural networks

AU - Hodge, V J

AU - Lees, K J

AU - Austin, J L

N1 - Copyright © 2003 Elsevier Ltd. This is an author produced version of a paper published in Neural Networks. This paper has been peer-reviewed but does not include the final publisher proof-corrections or journal pagination.

PY - 2004/4

Y1 - 2004/4

N2 - This paper evaluates a novel k-nearest neighbour (k-NN) classifier built from binary neural networks. The binary neural approach uses robust encoding to map standard ordinal, categorical and numeric data sets onto a binary neural network. The binary neural network uses high speed pattern matching to recall a candidate set of matching records, which are then processed by a conventional k-NN approach to determine the k-best matches. We compare various configurations of the binary approach to a conventional approach for memory overheads, training speed, retrieval speed and retrieval accuracy. We demonstrate the superior performance with respect to speed and memory requirements of the binary approach compared to the standard approach and we pinpoint the optimal configurations. (C) 2003 Elsevier Ltd. All rights reserved.

AB - This paper evaluates a novel k-nearest neighbour (k-NN) classifier built from binary neural networks. The binary neural approach uses robust encoding to map standard ordinal, categorical and numeric data sets onto a binary neural network. The binary neural network uses high speed pattern matching to recall a candidate set of matching records, which are then processed by a conventional k-NN approach to determine the k-best matches. We compare various configurations of the binary approach to a conventional approach for memory overheads, training speed, retrieval speed and retrieval accuracy. We demonstrate the superior performance with respect to speed and memory requirements of the binary approach compared to the standard approach and we pinpoint the optimal configurations. (C) 2003 Elsevier Ltd. All rights reserved.

KW - binary neural network

KW - associative memory

KW - correlation matrix memory

KW - k-nearest neighbour

KW - euclidean distance

KW - robust encoding

KW - quantisation

KW - binary mapping

U2 - 10.1016/j.neunet.2003.11.008

DO - 10.1016/j.neunet.2003.11.008

M3 - Article

SN - 0893-6080

VL - 17

SP - 441

EP - 458

JO - Neural Networks

JF - Neural Networks

IS - 3

ER -