Soft Classification Techniques for Breast Cancer Detection and Classification

Abstract

Breast cancer is known to be one of the most common cancers among women, often fatal. The reasons for death are mainly due to imprecision or delay in diagnosis. Early treatment helps to cure malignant growth and prevent its recurrence. The objective of this paper is to build a model to detect and correctly classify the tumor with high accuracy. In order to accomplish this, we compare Support Vector Machine (SVM) with five other Machine Learning (ML) Algorithms, namely, Decision Tree Classifier (CART), Naive Bayes Classifier (NB), Logistic Regression (LR), Linear Discriminant Analysis (LDA) and K Nearest Neighbor (KNN). ML Algorithms are known for their efficiency in data classification and are therefore widely used for diagnostic purposes in the medical field. We have evaluated the efficiency of SVM using precision, recall, ROC area and accuracy estimates. The best performance was achieved by the SVM method resulting in the highest accuracy.