Optimization of Breast Cancer Diagnostic Accuracy based on its Features using an Adaptive Principal Component Approach

Breast cancer develops when unchecked cell growth and division lead to a swelling of muscle identified as a tumor. Experiencing a lump present in the breast, noticing a transformation in the size of the breast, or noticing abnormalities to the skin on the breasts are all indications of breast cancer. Early detection is made possible via mammograms. The high rate of undetected malignancies is caused by the challenge of mammography interpretation. Consequently, it’s typically necessary to get a second opinion to lower the rate of errors. Radiologists frequently employ automated technologies to easily interpret mammograms. This research work provides an algorithm for feature extraction based on PCA and LR as a classifier as the best method to improve the separation of benign from malignant tissue. Additionally, the PCA rule of thumb—namely, the Eigenvectors and Eigenvalues, as well as Explained Variance—is used to identify features. According to preliminary findings, this method can distinguish the healthy and affected individuals.