CPM Homework Banner
12-48.

Governments and security companies are coming to rely more heavily on facial recognition software to locate persons of interest.

Consider a hypothetical situation. Suppose that facial recognition software can accurately identify a person of the time, and suppose the suspect is among facial images available to a government agency. When the software makes a positive identification, what is the probability that it is not the suspect?

  1. Make a model for this situation.

    Use the model at right as a starting point.
    Complete it by including the probabilities.

    2 by 2 rectangle, columns labeled, person, rows labeled, Facial ID Software, with additional labels as follows: top edge, left, Suspect , top edge, right, Not suspect. Left edge, top, Correct, left edge, bottom, Incorrect. Interior: top left, a, top right, b, bottom left, c, bottom right, d.

    P(suspect) 1/200,000, P(ID is correct)
    P(not the suspect) ?, P(ID is not correct) ?

  2. If a person has been identified as the suspect, what is the probability that he or she is not actually the suspect?

    The desired outcome is a person who is not the suspect being incorrectly identified, or situation D.
    The possible outcomes are situation D and the suspect being correctly identified, or situation A.