A type-2 neuro-fuzzy system with a novel learning method for Parkinson’s disease diagnosis

In this paper, an interpretable classifier using interval type-2 fuzzy neural network for detecting patients suffering from Parkinson's Disease (PD) based on analyzing the gait cycle is presented. The proposed method utilizes clinical features extracted vertical Ground Reaction Force (vGRF), measured by 16 wearable sensors placed in soles of subjects' shoes and learns rules. Therefore, experts can verify decision made investigating firing strength Moreover, utilize rules patient diagnosing or adjust them their knowledge. To improve robustness against uncertainty noisy sensor measurements, Interval Type-2 Fuzzy Logic applied. learn rules, two paradigms are proposed: 1- A batch learning approach clustering available samples applied to extract initial 2- complementary online rule base encountering new labeled samples. performance evaluated classifying healthy subjects different conditions including presence noise observing instances. model compared some previous supervised unsupervised machine approaches. final Accuracy, Precision, Recall, F1 Score 88.74%, 89.41%, 95.10%, 92.16%. Finally, sets each feature reported.