A Fuzzy Rough Model Using Genetic Algorithm for Accuracy Enhancement
International journal of intelligent engineering and systems · 2021
Fuzzy rules have been used extensively in data mining.Our model is proposed to construct Fuzzy Rules with considering criteria of accuracy.The model consists of five phases: (1) Automatic attributes fuzzification, (2) Feature selection via core and reduct, (3) Fuzzy rules extraction via SQL statement, (4) Calculates Accuracy and Confidence for each rule, (5) Genetic coding of fuzzy rules.The algorithm determines automatically the parameters width of each attribute then uses the rough set to reduce the number of attributes (Feature Selection) via core and reduct.The model then extracts fuzzy rules using SQL statements and calculates Accuracy and Confidence.Finally, our genetic model represents each fuzzy set by "Real number" to improve the accuracy.The model applied on Iris and Wine datasets and the result will be better than the other models in term of the number of fuzzy sets and classification rate for evaluating the accuracy of training and test instances.