By Bob Yirka , Tech Xplore
Article source: phys.org
Facebook has applied for a patent on technology the company has developed to collect data on people living in the same household by using photographs and other information stored or posted on the social networking site.
Facebook, like most other social network sites, not only keeps track of posting activity, but uses that data as a way to help sellers target users with ads more likely to get their attention. Now, it appears, engineers at Facebook have developed algorithms that target entire households using information from a variety of sources on the site, including posted pictures.
Such an algorithm could easily use information that is directly posted, such as home address, and names and descriptions of things that people do, along with information collected behind the scenes, such as IP addresses (multiple IP addresses from the same site give an indication of how many people live at a given location). It could also use physical GPS address location information collected from data stored along with digital images that users post from their phones.
The algorithm could also make use of reference data, such as groups of friends or family members mentioning attending the same event, or people appearing in photos posted by more than one person—like photos of a child by both a mother and a father. Also, it is likely that the algorithm would make use of keywords that appear in user posts, such as wife, sweetie, honey, etc.—clear indications of a familial relationship.