Skip to main content

View Diary: "Where were you on the night of June 10th, 2013?" (77 comments)

Comment Preferences

  •  I don't know what that means. Can you explain? (2+ / 0-)
    Recommended by:
    futurebird, mommyof3

    "Identify an individual from as few as 4 datapoints"

    Is it like Battleship? Or Karnak the Magnificent reading through an envelope?

    If someone was at 6th and Elm at 4:00 on Tuesday and Spruce and 11th on Thursday at 7:15 and ... and ... then: it MUST be Bill Miller!

    WHY??? I wanna play too! Tell me the secret!

    Please?

    Too late for the simple life, too early for android love slaves - Savio

    by Clem Yeobright on Mon Jun 10, 2013 at 04:36:34 PM PDT

    •  Here's my take on that - pretty much what you have (10+ / 0-)

      Scenario - unidentified individual was known to be:

      at location A at this date and time,
      location B at this date and time,
      location C at this date and time,
      location D at this date and time ...

      "Is the unidentified individual Clem Yeobright?:

      They will be correct 95% of the time. It's a matter of pattern-matching algorithms doing their magic.

      Free: The Authoritarians - all about those who follow strong leaders.

      by kbman on Mon Jun 10, 2013 at 04:52:50 PM PDT

      [ Parent ]

    •  Outside database is still needed (10+ / 0-)

      When I first read that passage, I assumed "individual" = a person identified by pattern matching, not necessarily the person's name, etc.

      The Nature article makes the link between my assumption and the personal identification:

      A simply anonymized dataset does not contain name, home address, phone number or other obvious identifier. Yet, if individual's patterns are unique enough, outside information can be used to link the data back to an individual.

      For instance, in one study, a medical database was successfully combined with a voters list to extract the health record of the governor of Massachusetts. In another, mobile phone data have been re-identified using users' top locations. Finally, part of the Netflix challenge dataset was re-identified using outside information from The Internet Movie Database.

      Here's an example as I see it in simpler terms.  There are three persons in my household with cell phones.  One goes to middle school regularly, one goes grocery shopping regularly, and one goes to Home Depot regularly.  Knowing through public records who the two adults residents are and assuming gender stereotypes, it's easy to match cell phone tracking to cell phone owner without even needing access to cell phone provider account records.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site