Related Experiment Videos
Enhancing heart disease prediction accuracy with hybrid machine learning.
Huie Zhang1,2, Caihong Li1,2, Xinzhi Tian1,2
1School of Electronic Information Engineering, Xi'an Siyuan University, Xi'an, Shaanxi, China.
Summary
This study enhanced cardiovascular disease prediction using hybrid machine learning models. The hybrid Histogram Gradient Boosting with Sea-Horse Optimizer (HGSH) model achieved the highest prediction accuracy, outperforming other models.
Area of Science:
- Medical Informatics
- Machine Learning
- Computational Biology
Background:
- Cardiovascular disease (CVD) poses a significant global health challenge.
- Accurate prediction models are crucial for early intervention and treatment.
- Existing machine learning models require optimization for improved predictive performance.
Purpose of the Study:
- To enhance cardiovascular disease prediction accuracy.
- To develop and evaluate hybrid machine learning models by integrating metaheuristic optimization algorithms.
- To compare the performance of novel hybrid models against baseline models.
Main Methods:
- Implemented adaptive boosting (ADA) and histogram gradient boosting (HGB) machine learning models.
- Integrated metaheuristic optimization algorithms: Sea-Horse Optimizer (SHO) and Chaos Game Optimizer (CGO).
- Developed hybrid models: ADSH (ADA + SHO), ADCG (ADA + CGO), HGSH (HGB + SHO), and HGCG (HGB + CGO).
Main Results:
- The HGSH model achieved the highest prediction accuracy (0.912).
- The HGCG model demonstrated strong performance with an accuracy of 0.902.
- Baseline ADA model showed lower precision (0.840) compared to hybrid models.
Conclusions:
- Hybrid models integrating metaheuristic optimizers significantly improve cardiovascular disease prediction.
- The HGSH model represents a superior approach for accurate CVD risk assessment.
- Optimized machine learning models offer a promising avenue for clinical decision support in cardiology.
