Heart attack prediction using artificial neural networks (ANN)
Code: G-1242
Authors: Azadeh Yaghoubi * ℗, احسان کوکائیان
Schedule: Not Scheduled!
Tag: Clinical Decision Support System
Download: Download Poster
Abstract:
Abstract
Background and aims: Cardiovascular diseases are emerging as a prominent health problem in modern societies. With the increase in risk factors such as unhealthy lifestyles, the need to identify and prevent heart attacks becomes more important. This paper examines the use of neural networks in predicting the risk of heart attack. Methods: For this research, the Python programming language and the TensorFlow library were used. The designed neural network, known as a feedforward multilayer perceptron (MLP), has one input layer, two hidden layers, and one output layer. The dataset used in this study includes information related to the cities of Hamedan, Nahavand, Malayer, and Tuyserkan from the hospital information system in the first 10 months of 1403. Among the measured variables are clinical and lifestyle factors, which ultimately provide the probability of the presence or absence of the risk of heart attack. Results: Using the collected data and applying the structured neural network model, an accuracy of 98.01% was achieved in diagnosing the risk of heart attack. These results indicate the high ability and reliability of neural networks in predicting heart problems. Conclusion: The present study shows that the use of neural networks, especially MLP models, can lead to significant improvements in heart attack prediction.
Keywords
Heart Attack, Neural Network, Perceptron
