Enhancement of performance and detection of lung disease using a novel grey level fuzzy neural network in comparison to Mamdani model fuzzy logic

The objective of this study is to enhance the detection accuracy of asthma disease in lungs by utilizing the Novel Gray level Fuzzy Neural Network (GFNN) with fuzzy rules, and to compare its performance with the Mamdani model Fuzzy Logic (FZ) algorithm. The study comprises two groups, with Group 1 employing the novel GFNN with fuzzy rules and Group 2 utilizing the Gray level Mamdani model Fuzzy Logic. Each group consists of 20 samples, determined using a pretest power of 80% and an error rate of 0.04, resulting in a total sample size of 40. The novel GFNN with fuzzy rules achieves a superior accuracy of 92.40% compared to the Mamdani model Fuzzy Logic, which achieves an accuracy of 90.30%. The obtained statistical significance value is 0.0213 (p<0.05), indicating a significant difference in performance between the two systems. Thus, the novel GFNN with fuzzy rules based feature extraction technique demonstrates superior accuracy in detecting asthma disease in lungs compared to the Mamdani model Fuzzy Logic based system.