Spam Analysis and Reputation Project

2. Standalone Approach

In Standalone approach, email messages from the existing mailboxes were used for analysis. A Graphical User Interface was used to perform tests. The user is asked to select and categorize the folders in the mailbox as spam folders, non spam folders and sent folders. Then all the message headers of the messages from the selected folders were retrieved using IMAP and the tests were performed on those messages to obtain corresponding results. The tests were performed on 22 different mailboxes from Gmail and Cubmail (Columbia University's Email Service) which consisted about 25,000 non spam messages and about 3000 spam messages.

Following are the statistics of the mailboxes that were used to perform tests:

MailBox #	MailBox	# of Mails	# of Non Spam mails	# of Spam mails
1	aditi_columbia	1818	1818	0
2	aditi_gmail	593	497	96
3	deepti_columbia	1174	1174	0
4	deepti_gmail	641	576	65
5	dhrumin_gmail	5105	5002	103
6	pinank_gmail	1682	1418	264
7	Preetinarayan_columbia	1230	1230	0
8	Preetinarayan_gmail	1992	1788	204
9	sneha_gmail	360	133	227
10	spinank_gmail	879	524	355
11	vasa_columbia	168	168	0
12	dms2169_columbia	1322	1301	21
13	nirav_gmail	1408	1360	48
14	nns_2108	934	934	0
15	manish_gmail	459	414	45
16	pragni_gmail	2183	1999	184
17	preetimalik_columbia	527	527	0
18	preetimalik_gmail	380	380	0
19	sak2144	749	749	0
20	shradha_columbia	140	140	0
21	shradha_gmail	1522	1151	371
22	vasa_gmail	3316	2370	946
Total		28582	25653	2929