Custom machine learning algorithm for large-scale disease screening - taking heart disease data as an example.

Area of Science:

• Cardiology
• Artificial Intelligence
• Public Health

Background:

• Heart disease is a leading global cause of mortality, necessitating effective large-scale screening strategies.
• Current screening methods face challenges in accuracy and scalability for widespread public health initiatives.

Purpose of the Study:

• To develop and evaluate a novel patient-specific machine learning algorithm for enhanced heart disease detection.
• To customize machine learning models by focusing on data processing, neural network architecture, and loss function formulation for improved accuracy.

Main Methods:

• Developed a patient-specific machine learning algorithm integrating individual patient data.
• Customized model development across data processing, neural network architecture, and loss function.
• Validated the algorithm using the Cleveland and UC Irvine (UCI) heart disease datasets.

Main Results:

• Achieved over 95% accuracy and recall on the Cleveland dataset.
• Exceeded 97% accuracy on the UCI dataset.
• Demonstrated superior performance in medical ethics and operability compared to general-purpose machine learning algorithms.

Conclusions:

• The patient-specific machine learning algorithm offers a powerful tool for effective large-scale heart disease screening.
• This approach has the potential to significantly improve patient outcomes and reduce the economic burden of heart disease.
• The customized model customization enhances reliability and applicability in real-world clinical settings.