Where will you be this time tomorrow? Smartphone data can guess within 20 metres -
Researchers have developed an algorithm that can predict where you will be in the future using data gathered from your friends smartphones.
In the study of two hundred people, the algorithm was able to predict the location of some users 24 hours later within 100 meters, some as close as 20 meters.
Mirco Musolesi, lead researcher and computer science lecturer at the University of Birmingham, said the algorithm is exploiting the synchronized rhythm of the city.
“I know this kind of thing looks kind of Big Brother,” said Musolesi.
“There is a problem for privacy, it is a big problem for these kind of services in general.”
Still, if users’ phone information could be protected, the algorithm has a potential future in targeted advertising, said Musolesi.
The algorithm predicts someone’s movements by comparing their data with people in their social group. For example, if Bob usually goes to the gym on Tuesday, but instead does an errand, the algorithm will look to his friends, Jack and Joe. If Jack and Joe are doing their normal routine, the algorithm will, probably correctly, predict Bob will continue on to the gym, his normal routine, following his errand.
“The mutual information between these two patterns is very high,” said Musolesi. “The movement of friends is highly coordinated, and synchronized.”
The Birmingham group won 3,000 euros for their work, part of a Nokia-sponsored Mobile Data Challenge, which gave a winning team an opportunity to analyze smart phone data —location, communication patterns and app usage— collected from 200 volunteers over the course of 18 months.
Read more -
http://www.thestar.com/business/article/1224211--where-will-you-be-this-time-tomorrow-smartphone-data-can-guess-within-20-metres
Researchers have developed an algorithm that can predict where you will be in the future using data gathered from your friends smartphones.
In the study of two hundred people, the algorithm was able to predict the location of some users 24 hours later within 100 meters, some as close as 20 meters.
Mirco Musolesi, lead researcher and computer science lecturer at the University of Birmingham, said the algorithm is exploiting the synchronized rhythm of the city.
“I know this kind of thing looks kind of Big Brother,” said Musolesi.
“There is a problem for privacy, it is a big problem for these kind of services in general.”
Still, if users’ phone information could be protected, the algorithm has a potential future in targeted advertising, said Musolesi.
The algorithm predicts someone’s movements by comparing their data with people in their social group. For example, if Bob usually goes to the gym on Tuesday, but instead does an errand, the algorithm will look to his friends, Jack and Joe. If Jack and Joe are doing their normal routine, the algorithm will, probably correctly, predict Bob will continue on to the gym, his normal routine, following his errand.
“The mutual information between these two patterns is very high,” said Musolesi. “The movement of friends is highly coordinated, and synchronized.”
The Birmingham group won 3,000 euros for their work, part of a Nokia-sponsored Mobile Data Challenge, which gave a winning team an opportunity to analyze smart phone data —location, communication patterns and app usage— collected from 200 volunteers over the course of 18 months.
Read more -
http://www.thestar.com/business/article/1224211--where-will-you-be-this-time-tomorrow-smartphone-data-can-guess-within-20-metres
No comments:
Post a Comment