A fuzzy logic based classification technique for clinical datasets

Data mining plays a vivacious role over the decision making condition under clinical data since huge volume of data are obtained day today form the medical field it becomes challenging to classify the data for each disease under certain criteria for the healthcare professionals. Ontology consists of hierarchical description of important classes in a particular domain; along with the description of the properties of each concept. In this paper a fuzzy ontology and fuzzy expert system consisting of fuzzification interface, fuzzy assessment K ratio algorithm and defuzzification interface are used. Using fuzzy ontology the data is solved effectively and handles knowledge with uncertainty and it includes fuzzy inference engine, fuzzy rule base for efficient classification of diabetes. The crisp data are converted into fuzzified data using fuzzification interface and defuzzification interface is used to convert the fuzzified value into crisp value. The fuzzy assessment K ratio is constructed using K ratio, correlation fuzzy logic and fact values to handle the uncertainty of rules. The K ratio technique is mainly used to overcome the overlapping of membership function by identifying the fuzzy mid value and fuzzy start value. The proficiency of the algorithm is carried out using the MATLAB Fuzzy logic toolbox to diagnosis diabetes.