Data Mining for Cardiovascular Disease Prediction

Area of Science:

• Computational intelligence
• Medical informatics
• Data science

Background:

• Cardiovascular diseases (CVDs) are a leading cause of global mortality and disability.
• Early identification of individuals at high risk for CVD is crucial for preventing premature death.
• Clinical data analysis using computational methods offers potential for improved CVD prediction.

Purpose of the Study:

• To apply Data Mining Techniques (DMTs) to clinical data for predicting cardiovascular diseases (CVDs).
• To evaluate the performance of various classification models in CVD risk assessment.

Main Methods:

• Utilized the CRossIndustry Standard Process for Data Mining (CRISP-DM) methodology.
• Applied five classifiers: Decision Tree (DT), Optimized DT, RIPPER (RI), Random Forest (RF), and Deep Learning (DL).
• Developed models using RapidMiner and WEKA tools, analyzing accuracy, precision, sensitivity, and specificity.

Main Results:

• The Optimized DT model demonstrated superior performance across all evaluation metrics.
• Optimized DT achieved 73.54% accuracy, 75.82% precision, 68.89% sensitivity, and 78.16% specificity.
• The Area Under the Curve (AUC) for the Optimized DT model was 0.788.

Conclusions:

• Data mining techniques show promise for effective CVD diagnosis.
• The Optimized DT model is a highly effective tool for predicting cardiovascular diseases.
• Further research into DMTs can enhance early CVD detection and patient outcomes.