Genetic algorithm fuzzy logic for medical knowledge-based pattern classification

No Thumbnail Available
Tan C.H.
Tan M.S.
Chang S.-W.
Yap K.S.
Yap H.J.
Wong S.Y.
Journal Title
Journal ISSN
Volume Title
Taylor's University
Research Projects
Organizational Units
Journal Issue
Hybrid of genetic algorithm and fuzzy logic in genetic fuzzy system exemplifies the advantage of best heuristic search with ease of understanding and interpretability. This research proposed an algorithm named Genetic Algorithm Fuzzy Logic (GAFL) with Pittsburg approach for rules learning and induction in genetic fuzzy system knowledge discovery. The proposed algorithm was applied and tested in four critical illness datasets in medical knowledge pattern classification. GAFL, with simplistic binary coding scheme using Pittsburg approach managed to exploit the potential of genetic fuzzy inference system with ease of comprehension in fuzzy rules induction in knowledge pattern recognition. The proposed algorithm was tested with three public available medical datasets, which are Wisconsin Breast Cancer (WBC) dataset, Pima Indian Diabetes dataset (PID), Parkinson Disease dataset (PD) and one locally collected oral cancer dataset. The results obtained showed that GAFL outperformed most of the other models that acknowledged from the previous studies. GAFL possessed the advantage of fuzzy rules extraction feature apart from conventional classification technique compared to other models which are lack of fuzzy interpretation. It is easier to interpret and understand fuzzy value in contrast to continuous or range value. GAFL outperformed the other algorithms in terms of accuracy without compromising on interpretability. It is vital to obtain high accuracy in medical pattern recognition especially when dealing with critical illness. � School of Engineering, Taylor�s University.