Spam Filtering Services

Spamming is overcrowding the internet with copies of the same message. The spam reaches the internet user even if the user does not choose to receive the message. Spamming is usually used for advertising as it is seen as a cheap way of promoting certain brands and goods. There are various types of spams which are:

E-mail Spam: The email spammers check every website for the '@' symbol to identify the email id. When the '@' symbol is encountered, the spammers collect the email ids and then send spam to all of them.

Virus Spam: If somebody else's computer is infected with computer virus then that virus picks up the e-mail addresses from the address book. The virus then sends spam to every email address in the book.

Even though, anti-spam laws are made, the spamming has not stopped. There are various ways to curb spamming, viz. filtering, blacklisting, white listing and using spamfree channels. The spam filters were initially located at the central servers. However, with the increase in the volume and the variety of spams, the spam filters are now installed on the local computers that are used. The spam filters are also installed at the central servers, but these spam filters filter the e-mails on the basis of content of the e-mail. The spam filters consist of classifiers, which can be signature based or rule-based. There are two types of spam filters which can be explained as follows:

Content or Rule-based Filters:
The rule-based filters follow certain rules to separate the spam e-mails. There are certain assumptions made, while designing the rule-based or content-based spam filters. These assumptions are as follows:

• The internet user always assumes that the unsolicited e-mails are spams.
• The e-mail that is posted several times is considered spam.
• All the internet users receive the same spam.

Nevertheless, people do not classify all the unasked or unsolicited e-mails as spam, which is a drawback of this spam filter. Also, the e-mails posted several times cannot be a spam, every time. More than one person can receive the same e-mail. For all these reasons, many spam filter makers have shifted to the Bayesian filtering.

Bayesian Filters (Learning filters)
The spam filters at the local machines (computer of internet user) do not filter the e-mails based on the content of the e-mail, but the rules are set by the filter software itself. These filters are called collaborative filters. When the user separates an e-mail as a spam, a signature is computed by the spam software and it is added to the database of spam filter. The next time the spam filter on the computer encounters that signature in some other mail, it is considered spam. The signature-based classifiers use the signature of the spammed e-mail to check for a spam every time.

Blacklist/ Whitelist Filtering
The blacklist/ whitelist filtering makes a list of spammers and also the list of emails which are solicited ones. Some filtering techniques are based on blacklist, some on white list and some on both. The blacklist filtering techniques use the signature-based filtering techniques to list out the spammed e-mails. The white list filters, on the other hand, make a list on their own. Every whitelist filter has its own list, based on the activities of the user.

Many spams do not get filtered even using spam filters. The spam filters are being developed further to provide good anti-spam services.