Monkey king evolution (MKE)-GA-SVM model for subtype classification of breast cancer

Area of Science:

• Oncology
• Machine Learning
• Bio-inspired Computing

Background:

• Breast cancer diagnosis relies heavily on accurate classification of clinicopathological data.
• Existing diagnostic tools can be improved with advanced computational approaches for better clinical decision-making.

Purpose of the Study:

• To develop an integrated metaheuristic machine learning model for precise breast cancer subgroup classification.
• To enhance the accuracy of breast cancer diagnosis using clinicopathological data.

Main Methods:

• Developed an integrated model combining Monkey King Evolution (MKE) for Support Vector Machine (SVM) hyperparameter tuning and Genetic Algorithm (GA) for feature selection.
• Compared the MKE-GA-SVM model against conventional methods like GA-SVM, grid search-SVM, and SVM-recursive feature elimination (RFE).
• Validated model performance using 10-fold cross-validation on three multicentre datasets and assessed using metrics like F1-score, ROC AUC, MSE, and logarithmic loss.

Main Results:

• The integrated MKE-GA-SVM model achieved superior classification accuracies of 91.4%, 86.6%, and 75.5% across three datasets.
• The developed model outperformed existing conventional feature selection and hyperparameter tuning methods.
• Comprehensive performance evaluation confirmed the model's effectiveness through various standard metrics.

Conclusions:

• The novel bio-inspired integrated metaheuristic model shows promise as a surrogate diagnostic tool.
• This advanced model can aid clinicians in making more precise diagnostic decisions.
• Deployment of this tool may lead to enhanced therapeutic outcomes for breast cancer patients.