Kernel PCA based on Hotelling Multivariate Control Chart for Monitoring Breast Cancer Diagnostic

Breast cancer is one of the diseases that is a scourge for women. The diagnosis of cancer is divided into benign breast cancer and malignant breast cancer. Accuracy in determining the diagnosis of breast cancer can help in patient treatment. To monitor breast cancer diagnoses, this article proposed a Kernel PCA based Hotelling’s T² Multivariate control chart. All features that are believed to affect cancer diagnosis are reduced with the Kernel PCA to overcome multicollinearity. The Kernel PCA based Hotelling’s T² The multivariate chart employs the bootstrap approach, a nonparametric resampling method to estimate the control limit. A study was conducted to compare the performance of the control charts with logistic regression to see the superiority of the control chart in diagnosing the type of breast cancer. The accuracy of Kernel PCA based Hotelling’s T² The multivariate graph is 89.63%. The logistic regression performance is better at classifying breast cancer diagnoses compared to Hotelling’s T² since it has a bigger accuracy. These results make sense because the function of logistic regression is to classify. Whereas in Hotelling’s T², we use the concept of in-control and out of control. However, to predict the diagnosis of breast cancer, the performance of Hotelling’s T² with an accuracy value close to 90%, can be said to be good.