Genetic Neuro-Fuzzy System for the Intelligent Recognition of Stroke

Abstract

Stroke is a global pandemic, affecting both developed and developing countries. In Nigeria, a steady rise in affected patients is becoming noticeable to all which inspired the development of this research. Stroke is caused by high blood pressure, smoking cigarettes, family history of stroke, high cholesterol, diabetes, obesity, overweight and cardiovascular diseases which affect the brain and damage part of the body (legs, hand) coordinated by that part of the brain. The symptoms of stroke vary from numbness of the affected body part to poor speech recognition and loss of balance. In this work, geno-neurofuzzy system for the intelligent recognition of stroke is designed. Genetic algorithm is used for optimizing fuzzy set or rules, neural network provides the self-learning paradigm while fuzzy logic handles vagueness or imprecision of fuzzy set. The evaluation results show an effective way of determining and assessing the three different levels of stroke. This provides a decision support for the tele-medical diagnosis of stroke within the health sector.

Keywords: Fuzzy logic, Genetic Algorithm, Neural Network, Stroke