Communications of the ACM
A Phone That Knows Where You're Going
Movement of three users around Switzerland's Lake Geneva are indicated with different symbols. GPS positions at the same time are indicated with the same color.
Credit: Mirco Musolesi / University of Birmingham
University of Birmingham researchers have developed an algorithm that follows a person's mobility patterns and adjusts for anomalies by also analyzing the patterns of people in the user's social network.
In a study of 200 volunteers, the system was an average of less than 20 meters off when it predicted where the users would be 24 hours later. However, the average error was 1,000 meters when the same system tried to predict a person's location using only their past movements and not those of their friends, according to Birmingham's Mirco Musolesi. He says the research is noteworthy because it exploits the synchronized rhythm of a city for greater predictive insights.
The research project was one of several at Nokia's Mobile Data Challenge. All of the projects drew on the same smartphone dataset from the same 200 volunteers. "It is exciting to see the project flesh out some of the hints and preliminary results we've seen in our earlier projects," says Massachusetts Institute of Technology researcher Alex Pentland. "This field is really moving toward being practical."
From Technology Review
View Full Article
Abstracts Copyright © 2012 Information Inc., Bethesda, Maryland, USA
No entries found