Fuzzy Logic-Based Inference System For Early Detection Of Cardiovascular Disease

Abstract

Cardiovascular disease is a primary cause of death. This research models and validates an innovative expert system using fuzzy inference to diagnose cardiovascular disease. The multi-layered fuzzy inference based on the Mamdani fuzzy model enables early classification of the disease, resulting in greater accuracy in the diagnosis depending on the class. The resultant medical expert system has the potential impact of saving lives. The system model comprises two layers. In layer 1, the input variables are blood pressure, diabetes, heredity, age, gender, and cholesterol. This layer detects the existence of the cardiovascular disease resulting in a binary Yes/No. Layer 2 input variables are low-density lipoprotein (LDL), high-density lipoprotein (HDL), Triglycerides, body mass index (BMI), Smoking, peripheral artery disease (PAD), and physical activity. The output variables of layer 2 are very low, low, medium, high, and very high.  Validation of the model evaluates its performance based on sensitivity, precision, specificity, classification accuracy, and F1 (96.06%, 93.15%, 96.06%, 95%, and 96.06%), respectively.