Spam filtering and identification is needed to stop Twitter becoming another platform for pushers of fake or substandard drugs, researchers say.
Promotion of fake pharmaceuticals via email is rife and Twitter presents counterfeiters with another platform to push their wares. A search for ‘buy AND Viagra’ by in-PharmaTechnologist found more than 10 spam messages in the past half-hour and the situation could get much worse.
“What we have found is that illegally pushing prescription drugs, or scamming people with sugar pills instead of legitimate drugs, fortunately has not caught on as rampantly in Twitter”, the University of Akron researchers wrote .
To restrict the rise of pharmaceutical spam on Twitter the researchers propose an identification and filtering system. Use of abbreviations, URL-shorteners and emoticons on Twitter, to convey meaning in 140 characters, create more problems in designing an effective anti-spam measure.
“Sites like Twitter and Facebook are growing so fast, it is difficult to determine which message or post is a spam and which is not without actually following the link”, the researchers wrote.
Identification systems based on 86 primary search terms, such as ‘Viagra’, and 45 secondary words, like ‘generic’, was created. Using the systems the researchers analysed two sets of Twitter posts and found a model using Random Tree learning algorithm and keyword-based decision logic worked best.
The nature of Twitter gives counterfeiters new “nefarious techniques to entrap victims”. Sending messages to masses of Twitter users and attaching popular search terms to spam posts are methods of getting people to visit fake online pharmacies.
“A favourite technique of spammers is to post tips on health, beauty, weight loss etc which are actually links to fraud websites. Self-promotional posts invite people to join communities, organizations and get free dining offers and other gifts”, the researchers said.