An Expert System for Anemia Diagnosis

The experience and knowledge accumulated making research on iron metabolism and erythroid disorders compelled us to tackle one of the most important clinical problems in hematology: anemia. The problem of unraveling the ethiology of an anemic state did not provide a fertile application for formal decision making and computer aided methods. Over the past three decades few researchers have applied statistical, pattern recognition and clustering techniques for building consultation models [1,2,4,8]. These techniques require large clinical data bases and have performed well at their developmental sites. Many problems limit a wider use of such systems [3,6].