Evaluation of RECIDEpsr - An Adviser in the Domain of Psychology

This paper introduces RECIDEPsY a casebased adviser system in the domain of psychology. RECIDEpsr has a library of success and failure cases. Cases of failure are in the form of indivisible and incompatible cases and represent constraints in the generation of a new solution from successful cases in memory. We present the experimental results obtained with RECIDEpsy. These results show that indivisible and incompatible cases reduce the number of successful cases necessary to cover the domain and decrease the number of wrong solutions given by the system.