Predictive Analysis of Breast Cancer Based on Stacking Algorithm

With the development of computers, machine learning algorithms can be applied in the medical field to solve many classification and prediction problems, thus assisting professionals to quickly judge and diagnose the disease. In this paper, we propose a breast cancer prediction model based on stacking algorithm, which integrates several traditional machine learning algorithms and compares with Adaboosting, SVM and other algorithms in terms of accuracy, ROC curve, PR curve, F1 value index, etc. The experiments show that the accuracy of the breast cancer classification model based on stacking algorithm can reach 97.23%, which is 6% higher than the classification accuracy of SVM, Adaboosting and other algorithms, and the AUC value of ROC curve can be improved by up to 0.26, which provides a certain reference value in breast cancer prediction examination and so on.