A Cloud-Based Predictive Model for the Detection of Breast Cancer

Abstract

Invasive cancer is the biggest cause of death worldwide, especially among women. Early cancer detection is vital to health. Early identification of breast cancer improves prognosis and survival odds by allowing for timely clinical therapy. For accurate cancer prediction, machine learning requires quick analytics and feature extraction. Cloud-based machine learning is vital for illness diagnosis in rural areas with few medical facilities. In this research, random forests, logistic regression, decision trees, and SVM are employed, and the authors assess the performance of various algorithms using confusion measures and AUROC to choose the best machine learning model for breast cancer prediction. Precision, recall, accuracy, and specificity are used to calculate results. Confusion matrix is based on predicted cases. The ML model's performance is evaluated. For simulation, the authors used the Wisconsin Dataset of Breast Cancer (WDBC). Through experiments, it can be seen that the SVM model reached 98.24% accuracy with an AUC of 0.993, while the logistic regression achieved 94.54% accuracy with an AUC of 0.998.