RANSAC Multi Model Response Regression based Pruned Extreme Learning Machines for Multiclass Problems

Lavneet Singh, Girija CHETTY

Research output: Contribution to journalArticle

Abstract

The accuracy and performance of machine learning and statistical models are still based on tuning some parameters and optimization for generating better predictive models of learning is based on training data. Larger datasets and
samples are also problematic, due to increase in computational times, complexity and bad generalization due to manually tuning of parameters. Using the motivation from extreme learning machine (ELM), we proposed annular ELM
based on RANSAC Multi Response Regression to prune the large number of hidden nodes to acquire better optimality, generalization and classification accuracy of the network in ELM. Experimental results on different benchmark
datasets showed that proposed algorithm optimally prunes the hidden nodes, better generalization and higher testing accuracy compared to other algorithms, including SVM, OP-ELM for binary and multi-class classification and regression problems.
Original languageEnglish
Pages (from-to)9-14
Number of pages6
JournalAustralian Journal of Intelligent Information Processing Systems
Volume14
Issue number2
Publication statusPublished - 2014

Fingerprint

Learning systems
Tuning
Testing

Cite this

@article{8909e0e95ae14ebbbd0a831fa0cdce31,
title = "RANSAC Multi Model Response Regression based Pruned Extreme Learning Machines for Multiclass Problems",
abstract = "The accuracy and performance of machine learning and statistical models are still based on tuning some parameters and optimization for generating better predictive models of learning is based on training data. Larger datasets andsamples are also problematic, due to increase in computational times, complexity and bad generalization due to manually tuning of parameters. Using the motivation from extreme learning machine (ELM), we proposed annular ELMbased on RANSAC Multi Response Regression to prune the large number of hidden nodes to acquire better optimality, generalization and classification accuracy of the network in ELM. Experimental results on different benchmarkdatasets showed that proposed algorithm optimally prunes the hidden nodes, better generalization and higher testing accuracy compared to other algorithms, including SVM, OP-ELM for binary and multi-class classification and regression problems.",
keywords = "optimized extreme learning machine, RANSAC Regularization, Multiclass-Problems",
author = "Lavneet Singh and Girija CHETTY",
year = "2014",
language = "English",
volume = "14",
pages = "9--14",
journal = "Australian Journal of Intelligent Information Proces...",
issn = "1321-2133",
number = "2",

}

TY - JOUR

T1 - RANSAC Multi Model Response Regression based Pruned Extreme Learning Machines for Multiclass Problems

AU - Singh, Lavneet

AU - CHETTY, Girija

PY - 2014

Y1 - 2014

N2 - The accuracy and performance of machine learning and statistical models are still based on tuning some parameters and optimization for generating better predictive models of learning is based on training data. Larger datasets andsamples are also problematic, due to increase in computational times, complexity and bad generalization due to manually tuning of parameters. Using the motivation from extreme learning machine (ELM), we proposed annular ELMbased on RANSAC Multi Response Regression to prune the large number of hidden nodes to acquire better optimality, generalization and classification accuracy of the network in ELM. Experimental results on different benchmarkdatasets showed that proposed algorithm optimally prunes the hidden nodes, better generalization and higher testing accuracy compared to other algorithms, including SVM, OP-ELM for binary and multi-class classification and regression problems.

AB - The accuracy and performance of machine learning and statistical models are still based on tuning some parameters and optimization for generating better predictive models of learning is based on training data. Larger datasets andsamples are also problematic, due to increase in computational times, complexity and bad generalization due to manually tuning of parameters. Using the motivation from extreme learning machine (ELM), we proposed annular ELMbased on RANSAC Multi Response Regression to prune the large number of hidden nodes to acquire better optimality, generalization and classification accuracy of the network in ELM. Experimental results on different benchmarkdatasets showed that proposed algorithm optimally prunes the hidden nodes, better generalization and higher testing accuracy compared to other algorithms, including SVM, OP-ELM for binary and multi-class classification and regression problems.

KW - optimized extreme learning machine

KW - RANSAC Regularization

KW - Multiclass-Problems

M3 - Article

VL - 14

SP - 9

EP - 14

JO - Australian Journal of Intelligent Information Proces...

JF - Australian Journal of Intelligent Information Proces...

SN - 1321-2133

IS - 2

ER -