Bill posted: "Rich is right, Janice. In addition to the points Rich makes
below, a good model developer will constantly test his model against real
data to find where its predictions deviate from actual data. Once the
reasons for the deviation are understood, the model can be improved to
reduce discrepancies from actual data. Thus the model does not serve as a
backup for the data, but its development cycle is intimately connected
with the real data. One question a policymaker can ask to ascertain
whether due diligence is being given to model development, is "Has the
model's performance improved over time?"...."
John Polkinghorne uses the word "corrigible" to encompass the above. It
is a good word to describe one important aspect of science.
To unsubscribe, send a message to email@example.com with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Thu Feb 22 16:45:50 2007
This archive was generated by hypermail 2.1.8 : Thu Feb 22 2007 - 16:45:50 EST