Evaluation of Some Selected Breast Cancer Classification Algorithms in Nigeria

Abstract

Breast Cancer (BC) is a prevalent disease that affects mostly women in the world. According to the World Health Organization (WHO), BC represent about 25 percent of all cancers in women with 685 000 deaths in 2020. An early detection of this disease can greatly increase the chances of taking the right decision on a successful treatment plan. This resulted in the need of new research avenues most especially in a country like Nigeria where there is low awareness of the disease and late presentation of BC by patients is normal. To achieve this, Support Vector Machine (SVM), KN Neighbor (KNN) and Decision Tree (DT) was used on a local dataset obtained from Ahmadu Bello University Teaching Hospital Zaria to provide some effective diagnostic capabilities. The dataset was classified into three classes (Benign, Pre-malign and Malign) and the SVM obtained a good classification accuracy of (99.2%). Late presentation of breast cancer is normal because of low awareness of the disease in the country therefore more awareness of the disease is highly recommended and women above the age of 34 years should always go for the breast cancer screening at least once a year with or without sign, sickness or symptoms.