Bina Estherly, Budi Prasetiyo
Breast cancer has become the most common cancer among women. In 2020 breast cancer ranks as the 5th cause of death from cancer and the number will continue to increase from year to year. In the development of information technology in health sector in the current era, diagnosis of a disease can be made based on existing historical data. One method that can be used is classification with data mining. In this study, the dataset used is the Original Wisconsin Breast Cancer Dataset obtained from the UCI Machine Learning Repository. The purpose of this study is to apply C4.5 algorithm with information gain as feature selection to select the most relevant attributes and the bagging ensemble method to overcome class imbalance in diagnosing breast cancer. For validate the performance of the models in this study uses 10-k fold cross validation and the accuracy results using the confusion matrix. The average accuracy produced is 96.49% while the accuracy results using the C4.5 algorithm without using information gain and bagging ensemble is 94, 58%. © 2023 Author(s).
Department of Computer Science, Universitas Negeri Semarang, Jalan Raya Sekaran Semarang, Central Java, Indonesia
