A New Machine Learning Technique for Breast Cancer Detection

Cancer, also called malignancy, is an abnormal growth of cells. it is characterized as a heterogeneous disease containing many various subtypes. The early diagnosis of cancer has a significant effect on the chance that the patient being cured. Automatically classifying cancer disease into two categories including high risk and low risk has been attracted many researchers in the fields of biomedical and bioinformatics. Recently, Machine Learning (ML) methods have been widely utilized in disease diagnosis. A variety of these techniques, including Articial Neural Networks (ANNs), Naive Bayes (NB), Support Vector Machines (SVMs), etc. have been widely applied in cancer detection and have achieved good results. In this paper, we review the most recent ML techniques in breast cancer research area and propose a new classifier based on a new loss function to automatically detect breast cancer and classify it into two categories. We investigate the effectiveness of our method through comprehensive experiments on several benchmark datasets. The experiments show an improvement compared with the state-of-the-art methods.

کلیدواژه ها:

component ، formatting ، style ، styling ، insert (key words)