A Fast Model Based on Genetic Algorithm to Construct Fuzzy Rules
Mohamed S.S. Basyoni
Mohamed S.S.Basyoni *, Department of Computer Sciences, Faculty of Graduate Studies for Statistical Research, Cairo University, Egypt.
Manuscript received on May 25, 2020. | Revised Manuscript received on June 29, 2020. | Manuscript published on July 30, 2020. | PP: 449-456 | Volume-9 Issue-2, July 2020. | Retrieval Number: B3592079220/2020©BEIESP | DOI: 10.35940/ijrte.B3592.079220
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Fuzzy rule has been used extensively in data mining. This paper presents a fast and flexible method based on genetic algorithm to construct fuzzy decision rule with considering criteria of accuracy. First, the algorithm determines the width that divides each attribute into “n” intervals according to the number of fuzzy sets, after that calculates the parameters width according to that width. Rough Sets Model Based on Database Systems technique used to reduce the number of attributes if there exists then we use the algorithm for extracting initial fuzzy rules from fuzzy table using SQL statements with a smaller number of rules than the other models without needing to use a genetic algorithm – Based Rule Selection approach to select a small number of significant rules, then it calculates their accuracy and the confidence.. Multiobjective evolutionary algorithms (EAs) that use nondominated sorting and sharing have been criticized mainly for computational complexity and needing for specifying a sharing parameter but in our genetic model each fuzzy set represented by “Real number” from 0 to 9 forming a gene on chromosome (individual). Our genetic model is used to improve the accuracy of the initial rules and calculates the accuracy of the new rules again which be higher than the old rules The proposed approach is applied on the Iris dataset and the results compared with other models: Preselection with niches, ENORA and NSGA to show its validity.
Keywords: Genetic algorithm, Fuzzy logic, Rough set, SQL statements, Accuracy.