Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data

Amongst all the machine learning techniques, kernel methods are increasingly becoming popular due to their efficiency, accuracy and ability to handle high-dimensional data. The fundamental problem related to these learning techniques is the selection of the kernel function. Therefore, learning th...

Full description

Saved in:

Bibliographic Details
Main Author:	Abbasnejad, M. Ehsan
Format:	Thesis
Language:	English
Published:	2010
Subjects:	QA75.5-76.95 Electronic computers Computer science
Online Access:	http://eprints.usm.my/41234/1/M._Ehsan_Abbasnejad-shahfiq.pdf
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my-usm-ep.41234
record_format	uketd_dc
spelling	my-usm-ep.412342018-08-06T07:45:15Z Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data 2010-05 Abbasnejad, M. Ehsan QA75.5-76.95 Electronic computers. Computer science Amongst all the machine learning techniques, kernel methods are increasingly becoming popular due to their efficiency, accuracy and ability to handle high-dimensional data. The fundamental problem related to these learning techniques is the selection of the kernel function. Therefore, learning the kernel as a procedure in which the kernel function is selected for a particular dataset is highly important. In this thesis, two approaches to learn the kernel function are proposed: transferred learning of the kernel and an unsupervised approach to learn the kernel. The first approach uses transferred knowledge from unlabeled data to cope with situations where training examples are scarce. Unlabeled data is used in conjunction with labeled data to construct an optimized kernel using Fisher discriminant analysis and maximum mean discrepancy. The accuracy of classification which indicates the number of correctly predicted test examples from the base kernels and the optimized kernel are compared in two datasets involving satellite images and synthetic data where proposed approach produces better results. The second approach is an unsupervised method to learn a linear combination of kernel functions. 2010-05 Thesis http://eprints.usm.my/41234/ http://eprints.usm.my/41234/1/M._Ehsan_Abbasnejad-shahfiq.pdf application/pdf en public masters Universiti Sains Malaysia Pusat Pengajian Sains Komputer
institution	Universiti Sains Malaysia
collection	USM Institutional Repository
language	English
topic	QA75.5-76.95 Electronic computers Computer science
spellingShingle	QA75.5-76.95 Electronic computers Computer science Abbasnejad, M. Ehsan Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data
description	Amongst all the machine learning techniques, kernel methods are increasingly becoming popular due to their efficiency, accuracy and ability to handle high-dimensional data. The fundamental problem related to these learning techniques is the selection of the kernel function. Therefore, learning the kernel as a procedure in which the kernel function is selected for a particular dataset is highly important. In this thesis, two approaches to learn the kernel function are proposed: transferred learning of the kernel and an unsupervised approach to learn the kernel. The first approach uses transferred knowledge from unlabeled data to cope with situations where training examples are scarce. Unlabeled data is used in conjunction with labeled data to construct an optimized kernel using Fisher discriminant analysis and maximum mean discrepancy. The accuracy of classification which indicates the number of correctly predicted test examples from the base kernels and the optimized kernel are compared in two datasets involving satellite images and synthetic data where proposed approach produces better results. The second approach is an unsupervised method to learn a linear combination of kernel functions.
format	Thesis
qualification_level	Master's degree
author	Abbasnejad, M. Ehsan
author_facet	Abbasnejad, M. Ehsan
author_sort	Abbasnejad, M. Ehsan
title	Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data
title_short	Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data
title_full	Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data
title_fullStr	Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data
title_full_unstemmed	Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data
title_sort	learning and optimization of the kernel functions from insufficiently labeled data
granting_institution	Universiti Sains Malaysia
granting_department	Pusat Pengajian Sains Komputer
publishDate	2010
url	http://eprints.usm.my/41234/1/M._Ehsan_Abbasnejad-shahfiq.pdf
_version_	1747820892242575360

Learning And Optimization Of The Kernel Functions From Insufficiently Labeled Data

Similar Items