New York: Researchers have used 500 million tweets to develop algorithms which they claim can predict an individual's behaviour hours in advance.
A typical social media post reveals much about a person: when they posted, where they were. The newly developed context-based algorithms can learn their networking relationships and predict what they plan to do.
Some people are very careful about what data they give out, but the algorithms can work pretty well with anonymised data, researchers from Binghamton University said.
Usable predictions can be made more than 60 per cent of the time, if the right data are aggregated.
Researchers said the data isn't coming just from social media but also from other sources such as credit card transactions, monitored telephone calls, e-mail and GPS data.
The new tools can be applied to call and contact centers. These methods can fuse data from call centers, online chat and e-mail help desks, researchers said.
"What if you call a company and they know why you're calling before you call?" researchers said.
The findings are published in the journal Industrial Engineer.