This paper reviews our work in the area of knowledge discovery in databases
. We have developed a genetic algorithm based machine learning system which
was successfully applied to discover diagnostic knowledge from female urin
ary incontinence and vertigo data. Comparisons with other methods suggest t
hat genetic algorithms are a competitive method to discover knowledge. We a
lso discuss an expert system for the differential diagnosis of female urina
ry incontinence that utilises the discovered knowledge. In addition, recent
research of data pre-processing with outlier identification and removal is
presented.