A hybrid colony fuzzy system for analyzing diabetes microarray data

Treatment to diabetes using microarray data has gained much attention among the physician as it provides important information about pathological states as well as information that can lead to earlier diagnosis. But its high dimensional low sample nature poses a lot of difficulties when it is analyzed by hand and needs an automatic system. As against statistical and machine learning approaches, fuzzy expert system provides an understandable diagnostic system. An important issue in the design of fuzzy expert system is knowledge acquisition. This paper presents a hybrid colony algorithm to extract if-then rules and to form membership functions from diabetes microarray data. During the run, Ant Colony Optimization (ACO) is used to generate optimal rule set and Artificial Bee Colony (ABC) is used to evolve the points of membership function. Mutual Information is used for identification of informative genes. The performance of the proposed approach is evaluated using two diabetes microarray data sets. From the simulation study, it is found that the proposed approach generated an accurate fuzzy system with interpretable rules when compared with other approaches.