MOSTAFA FATHI GANJI
MOHAMMAD SANIEE ABADEH

In this paper, we present a fuzzy rule-base classification system to detection of diabetes disease, named DiabMiner. DiabMiner system generates a set of fuzzy classification rules from labeled data by using an ant colony optimization (ACO) algorithm. These rules are represented in linguistic forms that are easily interpreted and examined by users. Each input pattern maybe compatible (can classify by multiple rules) with several fuzzy rules. Therefore, a fuzzy inference engine is used which classifies the input patterns based on multiple ifthen rules voting method. The results reveal that DiabMiner outperforms several famous methods in classification accuracy for diabetes disease detection.

Fuzzy Classification, Medical data mining, Ant Colony Optimization, Pattern Recognition