2. a) Evaluate the filtering features of ML mail program, especially usability of the filtering conditions. The user guide of ML is on http://www.cs.Helsinki.FI/compfac/mlmanual/manual.html; see Ch. 4 and especially Ch. 4.3. Give some examples of filter definitions (other than those in the manual ...). Considerations in point b (below) may be helpful in the evaluation of ML filtering.
b) Consider generally situations where you by filtering try to separate
certain mail messages: group messages from personal messages; reply
messages you are waiting for; a message which is new (unrelated to the
previous messages); messages needing fast response; important messages
(in some sense). What other separation needs there might exist?
Do the messages have a proper structure regarding these needs or
would it be possible to add some structural characteristics to make
filtering easier?
3. What are the main features of the Inquirus system presented in [2]?
References:
1. Amento, B., Terveen, L. & Hill, W., Does "authority" mean quality? Predicting expert quality ratings of Web documents. SIGIR 2000 , 296-303. (a copy in the course folder in A142)
2. Glover, E.J. et al., Recommending Web documents based on user preferences. (pdf; a copy in the course folder in A142)