Experiment for Programming 3B:
Purpose: Automatic identity recognition using the probabilistic conceptual approach we
discussed in the class.
Context:
Consider the settings for the showcase explored in the beginning of the
semester again: Johnny,
Winnie, Manny, and Cathy use their
one-dimensional keyboard (as described in the handout) to
type the Biola vision
statement and because of typos in the process we have a collection of 8
text files A, B, C, D, E, F, G and H (all updated Sept. 3). Each
of the eight documents was typed by one of Johnny, Winnie, Manny, and Cathy.
Approach: Apply
the probabilistic
conceptual approach to determine for each of the documents A, B, C, D, E, F, G and H the most likely person who has
generated the document when trying to type the Biola vision statement.
i.
What to do: (a) First, use option D provided by the demo
executable to determine log Pr(d |p) for
each document d and each person p. Record all the numerical results you
got and determine the most likely author for each document d. (b) Second, do the same things again except
that this time use your own program for Programming#3B to determine log Pr(d |p) for
each document d and each person p. Record all the numerical results you
got and determine the most likely author for each document d.
ii.
What to report: Record in the self-evaluation report
for Programming 3B all the numerical results and the answers about the most
likely author for each document in (a) and (b) above.