A CNN and PCA Approach for Earlier Breast Cancer Diagnosis
Hasna Elalaoui Eladallaoui, Adelaziz ElFazziki, Mohamed Sadgal · 2023
Breast cancer is one of the most harmful cancers in women. Breast cancer starts in the cells of the breast. Cancerous (malignant) tumor is a group of cancer cells that can invade and destroy nearby tissues. If cancer is not diagnosed when the first symptoms appear, it can spread (metastasize) to other parts of the body, causing other so-called later symptoms that are difficult to cure. It is in this context that machine learning applications have played a major role. Thanks to predictive models, we can detect the presence of a malignant tumor as early as possible and thus provide effective support in the clinical decision. Deep learning techniques can overcome the limitations of traditional algorithms by offering a high level of performance. This paper proposes a hybrid approach: unsupervised and supervised. The first is based on a descriptive analysis based on principal component analysis to understand the data available. The second is based on predictive analysis based on a convolutional neural network for the classification of benign and malignant massive tumors based on the characteristics of mammography images whose characteristics are archived in a dataset.