An Intelligent Approach for Early and Accurate Predication of Cardiac Disease Using Hybrid Artificial Intelligence...

Area of Science:

• Cardiology
• Machine Learning
• Artificial Intelligence

Background:

• Early and accurate diagnosis of heart disease is crucial for effective treatment and reducing mortality.
• Traditional machine learning and deep learning models have shown potential but can be further improved for diagnostic accuracy.

Purpose of the Study:

• To propose and evaluate a novel hybrid machine learning model, ETCXGB, for enhanced early diagnosis of heart disease.
• To compare the performance of the proposed ETCXGB model against other hybrid deep learning models and classical machine learning approaches.

Main Methods:

• Developed a hybrid model (ETCXGB) by combining ExtraTreeClassifier (ETC) and XGBoost (XGB) ensemble methods.
• Utilized predicted probabilities from ETC as input features for the XGB model to create an enriched feature matrix.
• Investigated other hybrid deep learning models: CNN-RNN, CNN-LSTM, and CNN-BLSTM for comparative analysis.

Main Results:

• The proposed ETCXGB model achieved a significant improvement in heart disease diagnosis accuracy.
• ETCXGB enhanced prediction accuracy by 3.91%, outperforming CNN-RNN (1.95%), CNN-LSTM (2.44%), and CNN-BLSTM (2.45%).
• Simulation outcomes confirmed that ETCXGB surpassed classical ML and DL models across all performance metrics.

Conclusions:

• The ETCXGB hybrid ML model demonstrates superior performance for accurate heart disease diagnosis.
• This advanced model can aid medical practitioners in making timely and precise diagnoses, potentially lowering cardiac disease mortality rates.