Using Machine Learning to Classify Heart Arrhythmias
Session Number
Project ID: : CMPS 02
Advisor(s)
Dr. Phadmakar Patankar; Illinois Mathematics and Science Academy
Discipline
Computer Science
Start Date
19-4-2023 10:20 AM
End Date
19-4-2023 10:35 AM
Abstract
The goal of the study is to research current technology being used to monitor heart rate patterns and learn how to incorporate machine learning to make current methods more effective and cost efficient.
This study utilizes a dataset containing records of patients’ age, weight, height, etc. along with typical electrocardiogram (ECG) test data. The first analysis approach involved the use of python clustering, categorizing the patients into two clusters to help distinguish between healthy patients and those afflicted with some type of arrhythmia based on the results of their ECG.
The clustering technique was first performed on the entire dataset, and then performed along with Principal Component Analysis (PCA) in order to reduce the dimensionality of the dataset. Reducing the dimensionality promotes computational efficiency while also maintaining the dataset’s variability when creating clusters.
The second approach utilizes neural networks, first without PCA and then with PCA, to categorize the patients according to the presence of arrhythmia. Upon the completion of both approaches, the study compares the results and corroborates the methods.
Using Machine Learning to Classify Heart Arrhythmias
The goal of the study is to research current technology being used to monitor heart rate patterns and learn how to incorporate machine learning to make current methods more effective and cost efficient.
This study utilizes a dataset containing records of patients’ age, weight, height, etc. along with typical electrocardiogram (ECG) test data. The first analysis approach involved the use of python clustering, categorizing the patients into two clusters to help distinguish between healthy patients and those afflicted with some type of arrhythmia based on the results of their ECG.
The clustering technique was first performed on the entire dataset, and then performed along with Principal Component Analysis (PCA) in order to reduce the dimensionality of the dataset. Reducing the dimensionality promotes computational efficiency while also maintaining the dataset’s variability when creating clusters.
The second approach utilizes neural networks, first without PCA and then with PCA, to categorize the patients according to the presence of arrhythmia. Upon the completion of both approaches, the study compares the results and corroborates the methods.