Collection Title:	SIU Thesis
Title :	A support vector machine approach to automatic email : multiclass classification
Material Type:	printed text
Authors:	Anmed, Mohamed Hussein, Author
Publisher:	Bangkok : Shinawatra University
Publication Date:	2014
Pagination:	vii, 43 p.
Size:	30 cm.
Price:	Gift.
General note:	Thesis. [M.S. [Information Technology]]. - Shinawatra University, 2014.
Languages :	English (eng)
Descriptors:	[LCSH]Electronic mail messages [LCSH]Electronic mail systems [LCSH]Electronic mail systems -- management [LCSH]Information retrieval [LCSH]Information Storage and Retrieval [LCSH]Machine learning [LCSH]WWW (Information retrieval system)
Keywords:	Information retrieval. Electronic mail system. Machine learning.
Class number:	SIU THE: SOIT-MSIT-2014-01
Abstract:	E-mail is a method of communication by sending and receiving digital messages among senders and recipients. Emails play a significant role in almost everyones’ daily activities. Typically users receive 20 to 30 or more messages per day and as the number of emails increase, users could unavoidably face an information overload which induces negative effects on the way users manage and organize their emails. Although, there are number of tools available for manual email management which focus on filtering messages by keyword, sender, etc., many of them are not very effective and troublesome to setup for novice users. This study focuses on email classification through grouping of messages based on their subjects and contents. The classification algorithms: the Supported Vector Machines (SVM) and Naïve Bayesian (NB) with kernel machines are studies and investigated. The study results reveal that SVM has a better performance than NB in both accuracy level and total computational time.
Curricular :	MSIT
Record link:	http://libsearch.siu.ac.th/siu/opac_css/index.php?lvl=notice_display&id=24134

SIU Thesis. A support vector machine approach to automatic email : multiclass classification [printed text] / Anmed, Mohamed Hussein, Author . - Bangkok : Shinawatra University, 2014 . - vii, 43 p. ; 30 cm.
Gift.
Thesis. [M.S. [Information Technology]]. - Shinawatra University, 2014.
Languages : English (eng)

Descriptors:	[LCSH]Electronic mail messages [LCSH]Electronic mail systems [LCSH]Electronic mail systems -- management [LCSH]Information retrieval [LCSH]Information Storage and Retrieval [LCSH]Machine learning [LCSH]WWW (Information retrieval system)
Keywords:	Information retrieval. Electronic mail system. Machine learning.
Class number:	SIU THE: SOIT-MSIT-2014-01
Abstract:	E-mail is a method of communication by sending and receiving digital messages among senders and recipients. Emails play a significant role in almost everyones’ daily activities. Typically users receive 20 to 30 or more messages per day and as the number of emails increase, users could unavoidably face an information overload which induces negative effects on the way users manage and organize their emails. Although, there are number of tools available for manual email management which focus on filtering messages by keyword, sender, etc., many of them are not very effective and troublesome to setup for novice users. This study focuses on email classification through grouping of messages based on their subjects and contents. The classification algorithms: the Supported Vector Machines (SVM) and Naïve Bayesian (NB) with kernel machines are studies and investigated. The study results reveal that SVM has a better performance than NB in both accuracy level and total computational time.
Curricular :	MSIT
Record link:	http://libsearch.siu.ac.th/siu/opac_css/index.php?lvl=notice_display&id=24134