The project aims to provide additional protection from unwanted, unsolicited bulk messages to email users by developing a spam filter that will able to learn patterns of spam based on the preferences of the user. The system will be using decision trees to classify whether a particular message is spam by recognizing patterns contained within the email messages which the user deems as spam.
Nowadays,
spammers are continually adapting new strategies on how they can fool existing
spam classifying softwares. The algorithm we are developing would be able
to adapt to new formats spammers use to trick other filters through its ability
to learn from the preferences of the user.
"I'm
sick and tired of getting spam mails", said student Adrian Lauron. Such
reaction is normal for a person who gets annoyed to unsolicited emails as
they clutter mailboxes. Irritation is not the only damage spam mails can inflict
to users; they can also prolong dial-up connections, wastes bandwidth, and
often expose minors to unsuitable contents.
The
growing volume of spam mails has challenged programmers to produce anti-spam
filtering mostly using Text Categorization techniques to automatically induced
UBE filters.