It has been argued that the MIN and MAX operators are not approptiate for the evulation of the AND and OR operators, respectively. This is because the MIN and MAX operators do not well agree with human's behavior for document ranking. There is a variety of fuzzy operators to overcome the problem of the MIN and MAX operators. In this thesis we first address important issues to affect retrieval effectiveness, which is based on the behavioral aspects of the fuzzy operators. We then present that Information Retrieval systems achieve high performance using a class of fuzzy operators called positively compensatory operators. We have also shown through performance experiments that positively compensatory operators provide better retrieval effectiveness than the others.
