Heart Attack Prediction System Using IoT and Machine Learning

Abstract

Abstract: Nowadays, health diseases are increasing day by day due to lifestyle, hereditary. Especially heart attack has become more common lately, i.e., the life of people is at risk. Each individual has different values for vital sign, cholesterol, and pulse. But consistent with medically proven results the traditional values of vital sign are 120/90, Cholesterol is 100- 129 mg/dL, pulse is 72, Fasting blood glucose level is 100 mg/dL, Heart rate is 60-100bpm, ECG is normal, Width of major vessels is 25 mm (1 inch) in the aorta to only 8 m in the capillaries. This paper analyses various classification systems for determining a person's risk level based on age, gender, blood pressure, cholesterol, and pulse rate. A predictive modelling-based "Ailment Prediction" system predicts the user's disease based on the symptoms provided as input to the system. The system takes the input from user and analyses the symptoms and gives the output as a probability of the disease. Five approaches are used to predict disease: Naive Bayes, KNN, Decision Tree, Linear Regression, and Random Forest Algorithms. These methods are used to evaluate the disease's probability. Therefore, the average prediction accuracy probability of 83 % is obtained. Keywords: Heart rate sensor Pulse, Android smartphone, Pulse Sensors, ECG sensor, Internet of Things.