An Image Processing Approach for Detection of Prenatal Heart Disease

Area of Science:

• Medical Informatics
• Cardiology
• Data Science

Background:

• Congenital heart defects (CHDs) are a leading cause of global mortality.
• Accurate early detection of cardiovascular disease (CVD) risks is crucial.
• Data preprocessing and mining are vital for analyzing complex medical data.

Purpose of the Study:

• To develop a predictive model for cardiovascular disease.
• To evaluate the effectiveness of various data mining classification algorithms for CVD risk prediction.
• To identify key attributes for accurate CVD forecasting.

Main Methods:

• Utilized a dataset of 300 patients with heart disease from a medical research database.
• Selected 15 relevant attributes from an initial set of 75 for analysis.
• Employed classification models including Naïve Bayes (NB), K-Nearest Neighbors (KNN), Decision Tree (DT), and Random Forest (RF).

Main Results:

• The Naïve Bayes classifier achieved the highest overall accuracy in predicting cardiovascular disease.
• Analysis focused on 15 key variables to establish the utility of different predictive approaches.
• The study demonstrates the potential of data mining for early CVD risk identification.

Conclusions:

• Naïve Bayes is a highly effective algorithm for predicting cardiovascular disease risk.
• Data mining techniques offer a promising avenue for improving early detection and management of heart conditions.
• Further research can refine these models for clinical application in preventing CVD.