Predicting the Risk of Heart Attack Using Machine Learning Algorithms

Abstract

This chapter uses two datasets to predict heart attacks, which provide valuable insights into the machine learning algorithms used in the prediction process. The first dataset used in this chapter is the heart attack risk prediction dataset, which has 14 attributes relevant to heart health, including demographic information of the patient, clinical features, diagnostic test results, symptoms, risk factors, and outcome variables. The second dataset used in this chapter is the heart attack dataset collected at Zheen Hospital in Erbil, Iraq, in 2019. The dataset has 1319 records of patients and nine attributes indicating the heart health status of the patients. The categorical data is normalized and represented in the dataset. This chapter aims to predict the risk of a heart attack using three machine learning algorithms. Since both datasets used in this chapter have labeled data, these algorithms should fall under the category of supervised machine learning, where the machine learns from the data provided in the training process to predict the results for the test data. Decision Tree, Random Forest, and K Nearest Neighbors (KNN) are the methods selected for this chapter based on several criteria. Incorporating several datasets permits the identification of issues related to data heterogeneity and sample size limitations and, thereby, the examination of the model's generalizability and reliability.

Ähnliche Arbeiten