Imperfectly Separable Models
, by Jon F. Claerbout
Textbooks, particularly those on partial differential equations and those on
probability, are filled with separable models of reality. The real world
contains many data bases which are roughly separable. But poor agreement
between theory and practice can often be traced to sensitivity of results to
departure from perfect separability. This paper describes the problem in
general, shows an effective technique to deal with it, and then cites specific
examples.
STANFORD