Machine Learning Technology-Based Heart Disease Detection Models.

Area of Science:

• Cardiology
• Computer Science
• Artificial Intelligence

Background:

• Heart failure disease is a significant global health concern.
• Early detection of heart disease is critical for effective healthcare services.
• Electrocardiogram (ECG) is a standard diagnostic tool, but advanced methods are needed.

Purpose of the Study:

• To analyze various machine learning technologies for heart disease detection.
• To provide clinicians with a tool for early heart problem diagnosis.
• To improve the accuracy of heart disease diagnosis using machine learning.

Main Methods:

• Utilized Naïve Bayes with a weighted approach for heart disease prediction.
• Employed Support Vector Machine (SVM) and XGBoost for ischemic heart disease detection based on frequency and time domain features.
• Developed a heart failure prediction model using DBSCAN for outlier detection, SMOTE-ENN for data balancing, and XGBoost for classification.

Main Results:

• Compared four machine learning models based on precision, accuracy, f1-measure, and recall.
• XGBoost demonstrated strong performance in classification tasks for heart disease prediction.
• The study investigated alternative decision tree classification algorithms to enhance diagnostic accuracy.

Conclusions:

• Machine learning offers valuable tools for disease diagnosis, detection, and prediction in the medical industry.
• Early diagnosis through improved prediction models can lead to more effective patient treatment and prevention of severe complications.
• The integration of advanced machine learning techniques can significantly aid clinical decision support systems.