Density Estimation as a Preprocessing Step for Constructive Algorithms

Jörg C. Lemm; Valeriu Beiu; John G. Taylor

doi:10.1007/978-1-4471-3087-1_41

Density Estimation as a Preprocessing Step for Constructive Algorithms

Jörg C. Lemm, Valeriu Beiu, John G. Taylor

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

Abstract

The aim of the paper is to combine the generalization power of classical density estimation techniques with the efficiency of VLSI-friendly, constructive algorithms. This is important for VLSI-implementations of classification networks in the case of noisy data to avoid the well-known “overfitting” effects. The method consists of three steps 1) estimation of the probability densities for each class, 2) discretization of the input space and 3) describing the resulting classification regions using only easy implementable boolean AND and OR gates and comparisons. The “noisy spirals” classification problem, a noisy variant of the “two spirals” benchmark, is used for demonstration.

Original language	English
Title of host publication	Proceedings of the Third Annual SNN Symposium on Neural Networks
DOIs	https://doi.org/10.1007/978-1-4471-3087-1_41
Publication status	Published - Sept 14 1995
Event	SNN'95 - Nijmegen, The Netherlands Duration: Sept 14 1995 → …

Conference

Conference	SNN'95
Period	9/14/95 → …

Access to Document

10.1007/978-1-4471-3087-1_41

http://link.springer.com/chapter/10.1007%2F978-1-4471-3087-1_41

Cite this

@inproceedings{5740bcbb71b74fd9810a190738d1ba54,

title = "Density Estimation as a Preprocessing Step for Constructive Algorithms",

abstract = "The aim of the paper is to combine the generalization power of classical density estimation techniques with the efficiency of VLSI-friendly, constructive algorithms. This is important for VLSI-implementations of classification networks in the case of noisy data to avoid the well-known “overfitting” effects. The method consists of three steps 1) estimation of the probability densities for each class, 2) discretization of the input space and 3) describing the resulting classification regions using only easy implementable boolean AND and OR gates and comparisons. The “noisy spirals” classification problem, a noisy variant of the “two spirals” benchmark, is used for demonstration.",

author = "Lemm, {J{\"o}rg C.} and Valeriu Beiu and Taylor, {John G.}",

year = "1995",

month = sep,

day = "14",

doi = "10.1007/978-1-4471-3087-1_41",

language = "English",

isbn = "978-3-540-19992-2",

booktitle = "Proceedings of the Third Annual SNN Symposium on Neural Networks",

note = "SNN'95 ; Conference date: 14-09-1995",

}

TY - GEN

T1 - Density Estimation as a Preprocessing Step for Constructive Algorithms

AU - Lemm, Jörg C.

AU - Beiu, Valeriu

AU - Taylor, John G.

PY - 1995/9/14

Y1 - 1995/9/14

N2 - The aim of the paper is to combine the generalization power of classical density estimation techniques with the efficiency of VLSI-friendly, constructive algorithms. This is important for VLSI-implementations of classification networks in the case of noisy data to avoid the well-known “overfitting” effects. The method consists of three steps 1) estimation of the probability densities for each class, 2) discretization of the input space and 3) describing the resulting classification regions using only easy implementable boolean AND and OR gates and comparisons. The “noisy spirals” classification problem, a noisy variant of the “two spirals” benchmark, is used for demonstration.

AB - The aim of the paper is to combine the generalization power of classical density estimation techniques with the efficiency of VLSI-friendly, constructive algorithms. This is important for VLSI-implementations of classification networks in the case of noisy data to avoid the well-known “overfitting” effects. The method consists of three steps 1) estimation of the probability densities for each class, 2) discretization of the input space and 3) describing the resulting classification regions using only easy implementable boolean AND and OR gates and comparisons. The “noisy spirals” classification problem, a noisy variant of the “two spirals” benchmark, is used for demonstration.

U2 - 10.1007/978-1-4471-3087-1_41

DO - 10.1007/978-1-4471-3087-1_41

M3 - Conference contribution

SN - 978-3-540-19992-2

BT - Proceedings of the Third Annual SNN Symposium on Neural Networks

T2 - SNN'95

Y2 - 14 September 1995

ER -

Density Estimation as a Preprocessing Step for Constructive Algorithms

Abstract

Conference

Access to Document

Fingerprint

Cite this