At 08:32 PM 6/20/2003, Kathryn Blackmond Laskey wrote: >At 7:30 AM -0700 6/20/03, Marco Zaffalon wrote: > >...I am happy with systems that can recognize the limits > >of their knowledge and suspend the judgment when these limits are reached, > >in the same way that I prefer to be told "I do not know" when I ask for > >road information rather than being recommended a wrong route. > >Even better is a system that can say: "If forced to give advice, my >recommendation is X, with justification Y, but my confidence in that >recommendation and its justification is only Z." If the answer is not >very trustworthy, I would be well-advised to consult other sources if >resources permit, but at the very least I have a recommendation and a >justification to chew over. A system that says nothing but that it >suspends judgment gives me no help at all. > >Kathy
I would like to see a system of this kind say, in addition: "These advices and precise confidences have been obtained by doing the following assumptions. Are these assumptions acceptable to you?". The responsibility of doing key assumptions would be left to the user in this way. Furthermore, it is unlikely that the user always accepts the proposed assumptions (e.g., assuming that the data or the evidence are missing at random; assuming that prior ignorance can be satisfactorily modelled by a single prior density; etc...), so the user would be left, occasionally, with indeterminacy and imprecise probabilities. As a side comment, it may be worth emphasizing that "I do not know" (i.e, the output is the set of all the possible options) is only an extreme case of indeterminacy. For example, consider a problem of medical diagnosis with 10 possible disease. Given evidence on a patient, the system might exclude 8 diseases and output the remaining 2 (with related posterior intervals of probability - which would overlap in this case). This appears to be quite an informative output, despite being partially indeterminate. Marco