A hybrid breast cancer classification algorithm based on meta-learning and artificial neural networks

Area of Science:

• Oncology
• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Breast cancer is the leading new cancer case globally, surpassing lung cancer in women.
• Early detection through regular screening and checkups is crucial for effective treatment.
• Computer-aided technologies can enhance diagnostic accuracy and physician decision-making.

Purpose of the Study:

• To develop a hybrid algorithm integrating meta-learning and Artificial Neural Networks (ANN) for breast cancer prediction.
• To improve the accuracy and efficiency of breast cancer diagnosis using advanced computational methods.
• To provide a robust computational tool for early breast cancer detection.

Main Methods:

• Data pre-processing and feature selection using a correlation-based method, identifying 16 key features.
• Development of a hybrid algorithm combining meta-learning models with Artificial Neural Networks (ANN).
• Utilizing the output of meta-learning models as input features for the ANN models.

Main Results:

• The proposed hybrid algorithm achieved a prediction accuracy of 98.74%.
• An F1-score of 98.02% was obtained, indicating high precision and recall.
• The hybrid model demonstrated superior prediction performance compared to individual models.

Conclusions:

• The hybrid meta-learning and ANN algorithm offers a highly accurate and efficient method for breast cancer prediction.
• This computational approach can significantly aid in early detection and improve patient outcomes.
• The study highlights the potential of combining machine learning techniques for complex medical diagnoses.