There are lots of cancer types affects the well-being of human race. Among those breast cancers affected persons are rapidly increasing day by day. The medications and treatment methodologies followed for breast cancer are not sufficient enough to detect the infections in early stage. The prediction process followed in breast cancer detection only follows physical examination. This type of cancer is mostly found on female and less common in men, it’s very difficult to find the presence of cancer cells in early stage. The patients coming for treatment only after the pain on breast and different in breast. Mostly breast cancers are identified at the second stage only and remains difficult in treatment process in later stages. The focus of the research work is to identify the breast cancer in early stage before the infections affects more to second stage for better treatment process. The collected breast cancer data are preprocessed before taking into feature extraction process and analyzing stage. This paper examines different classification methods followed during preprocessing and measurement matrices followed for following supervised learning and semi supervised learning. This paper shows different methodologies implemented in preprocessing technique for removing irrelevant information and duplicates in extracted data from images

Discretization, Resampling, magnetic resonance imaging, Logistic Regression.