IBM Develops Algorithm to Predict Where You Live Based on Your Last 200 Tweets
IBM researchers have developed an algorithm they claim can determine your home location with 70% accuracy, based on your last 200 tweets.
In a paper written by Jalal Mahmud, Jeffrey Nichols and Clemens Drews of the IBM Research Lab in Almaden (PDF), the researchers say their algorithm “uses an ensemble of statistical and heuristic classifiers to predict locations and makes use of a geographic gazetteer dictionary to identify place name entities.”
Some people make it easy to determine their location, by using apps like FourSquare or city-based expressions (like “go Red Sox” for Boston). But other data, which allowed researchers to determine time zone based on the time of day of the tweets, may be more difficult for users to control.
The researchers say this kind of data allows them to determine when someone is traveling. And when they exclude travelers, they claim they can determine someone’s home city with 68 per cent accuracy, their home state 70 per cent of the time and their time zone 80 per cent of the time.
Thus people who are otherwise scrupulous about their privacy and the amount of information they allow to be scooped online up may still be giving away more data on themselves than they realize.
Facebook researchers also announced this week that they now have the ability determine someone’s identity 97.25% of the time in photos using facial recognition technology, even in random snapshots, using the nefarious-sounding “Deep Face” technology.
Bottom line: it’s getting ever more difficult to protect your privacy online. You’re probably giving up more information than you know.