KNN Algorithm Used for Heart Attack Detection

Abstract

Machine Learning, a branch of artificial intelligence, has become more accurate than human medical professionals in predicting the incidence of heart attack or death in patients at risk of coronary artery disease. In this paper, we attempt to employ Artificial Intelligence (AI) to predict heart attack. For this purpose, we employed the popular classification technique named the K-Nearest Neighbor (KNN) algorithm to predict the probability of having the Heart Attack (HA). The dataset used is the cardiovascular dataset available publicly on Kaggle, knowing that someone suffering from cardiovascular disease is likely to succumb to a heart attack. In this work, the research was conducted using two approaches. We use the KNN classifier for the first time, aided by using a correlation matrix to select the best features manually and faster computation, and then optimize the parameters with the K-fold cross-validation technique. This improvement led us to have an accuracy of 72.37% on the test set.

Smart K Nearest Neighbor Outlier Detection for Electroencephalogram signal

“…Data mining is widely used in the medical field to detect outliers then allow doctors to take the best decisions [12]. Several research efforts [3,30,31,32] in the past have used the KNN to address various challenges in the medical field in general and especially for BCI and EEG. KNN is one of the most popular algorithms applied to several data sets.…”

Perbandingan Penggunaan Jarak Manhattan, Jarak Euclid, dan Jarak Minkowski dalam Klasifikasi Menggunakan Metode KNN pada Data Iris

“…Penelitian tentang klasifikasi dengan metode KNN juga digunakan dalam penelitian (Ali et al, 2020;Bah & Yu, 2021;Pamuji, 2021;Pratama et al, 2021;Rudiyan et al, 2022;Wahyono et al, 2020;Yunus et al, 2021). Dalam penelitian (Wahyono et al, 2020) digunakan data uji 10 %, p = 5 dan k bervariasi dari 3 sampai dengan 9, namun hanya menggunakan satu nilai random state yaitu 0.…”

Machine Learning Techniques for Heart Disease Classification Using K-Nearest Neighbor Optimization with Particle Swarm Optimization

“…The K-fold cross-validation technique is then used to refine the parameters. We achieved an accuracy of 72.37% on the test set as a result of this modification [8]. The third research was conducted by [9], In this study, various values of k, specifically 1, 3, 5, 7, and 9, will be tested in order to determine the most appropriate value based on the accuracy achieved from calculations.…”