Crowdsourced software could stop SMS spam - tech - 26 February 2011 - New Scientist

Is your cellphone buzzing with unwanted text messages? A system that filters out SMS spam by enlisting the help of your friends could calm things down.

In the western world, receiving an illicit text is a minor annoyance, but it's a major problem in developing countries like India. In that country, it is estimated that 100 million spam messages are sent every day, according to a report issued last year by the Telecom Regulatory Authority of India.

"That was one of the big motivations for us to start looking at this problem," explains Ponnurangam Kumaraguru, who developed the software package, SMSAssassin, with colleagues at the Indraprastha Institute of Information Technology in New Delhi, India.

SMS filtering is not a new idea, and existing methods work much like email spam detectors. They all learn to identify spam messages by examining known spam, but this is less effective for SMS messages because their brevity makes it hard to identify features unique to spam. Abbreviations and regional words, common in text messages, make this even worse.

Kumaraguru's team worked around these limitations by relying on crowdsourced spam markers. SMSAssassin learns in the same way as other spam filters, and the researchers hope toone day allow users to share spam keywords with one another through a central server or by creating a distributed network via Bluetooth.

The team say this will let the system react quickly to new kinds of spam or messages tied to certain time periods, such as the Diwali religious festival. For now they are gathering user data by asking users to contribute their spam through a Facebook page.

Read the full post on newscientist.com

Posted via email from Siobhan O'Flynn's 1001 Tales