Multiclass classification of breast cancer large scale datasets for detecting cancer drivers

Abstract

Over the past decade, scientists have found that even healthy genes can cause cancer due to hormonal growth disorder. The pattern of multiclass classification on data mining has recently become an important topic for research, especially in the health sector. Classification of cancer cells also plays an important role in the development of almost all types of cancer, and in this case, we focus on breast cancer. Therefore, studying Multiclass Classification is crucial to the experts in diagnosing cancer. Since datasets on type of breast cancer cell are plenty, it is important to pay more attention to the method to be as efficient as it could be for we are going to process such large datasets. Based on big data technologies, this study proposes the feature selection step in high dimension data classification problem and datasets with dozens of features. Multiclass Classification supports a study to adopt big data solutions. This machine learning techniques analyze a breast mass by analyzing the digitized image of a fine needle aspirate (FNA) which describes characteristics of the cell nuclei present in breast cancer. From the datasets of various classifications of breast mass will be investigated further to determine their active role in cancer. Especially, based on this research aimed to identify and analyze the ability of Support Vector Machine (SVM) as a Classification method and Relief F-Based Feature Selection as a Selection Method for diagnosing breast cancer driver. This method could be an efficient method for cancer classification with the accurate performance of 91 %.