Klasifikasi Kanker Payudara Menggunakan Algoritma SVM dengan Kernel RBF, Linier, dan Sigmoid

Abstract

Breast cancer ranks first in both the gender category and the death rate. Late treatment is often found in cases of breast cancer which causes an increase in the risk factors for this cancer. For this reason, early detection of breast cancer is needed, so that treatment can be done in a timely manner, so that the death rate due to breast cancer can be reduced. For this reason, this article offers early detection of breast cancer using classification. The dataset in this study used the Wisconsin breast cancer dataset taken from Kaggle. Initially the dataset has a missing value, besides that the categorical data is not yet in numerical form, so it is necessary to do preprocessing with the missing value imputing technique and encoding to convert categorical data into numeric data. The dataset is divided into two proportions, namely 80% as training data and 20% as testing data. In the classification process, datasets that have been preprocessed are classified using SVM with three different kernels, namely the linear kernel, the RBF kernel, and the Sigmoid kernel. Based on the research results that have been obtained, the linear kernel shows the best classification results when applied to the SVM classification, namely with an accuracy value of up to 99%, followed by RBF kernel performance with an accuracy rate of 92%, and finally the sigmoid kernel with an accuracy value of 41%