Break the Ice (Opthygb): Hyper-Parameter Optimization Predictive Machine Learning Model Framework for Early Warning Breast Cancer

Abstract

Abstract— Breast cancer, characterized by abnormal cell growth in the breast, predominantly affects women. In the pursuit of more accurate prediction and diagnosis, researchers are turning to artificial intelligence. One widely used dataset for this purpose is the Wisconsin Breast Cancer Dataset (WBCD), offering 30 metrics, including mean, standard error, and 'worst' values, to shed light on breast cancer characteristics.In this study, we introduce a novel model named OPThyGB, integrating Gradient Boosting with the OPTUNA optimizer. The model's performance is evaluated using essential metrics such as ROC curve, AUC, specificity, F1-score, sensitivity, and accuracy. Impressively, the hyperparameter optimization technique results in an outstanding 98.25% accuracy, surpassing benchmarks set by previous studies.This achievement not only showcases the model's exceptional accuracy but also positions it as a frontrunner in breast cancer prediction when compared to existing literature. The study underscores the potential of machine learning algorithms to significantly enhance breast cancer prediction and diagnosis, thereby contributing to improved clinical decision-making and ultimately better patient outcomes.

Note:
Funding Declaration: This work was supported by the Industry- University-Research Innovation Fund for Chinese Universities (No.2021FNA04004), Open Research Fund from Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ)(No.GML- KF-22-07).

Conflicts of Interest: None.

Keywords: IndexTerms—Machinel learning, OPTUNA, Optimization, Breast cancer, Hyper-parameter