Automated discovery of positive and negative knowledge in clinical databases

Describes a rule-induction method based on rough-set models that more closely represents medical experts' reasoning. The characteristics of two measures, classification accuracy and coverage, are discussed. The author shows that both measures are dual, and that accuracy and coverage are measures of both positive and negative rules, respectively. Then, an algorithm for induction of positive and negative rules is introduced. The proposed method is evaluated on medical databases, and the experimental results show that induced rules correctly represent expert knowledge. Several interesting patterns are also discovered.