Breast cancer is the second leading cause of death due to cancer in women currently. It has become the most common cancer in recent years. In early detection of cancer, data mining can be used to diagnose breast cancer. Data mining consists of several research models, one of which is classification. The most commonly used method in classification is the decision tree. C4.5 is an algorithm in the decision tree that is often used in the classification process. In this study, the data used was the Breast Cancer Wisconsin (Original) Data Set (1992) obtained from the UCI Machine Learning Repository. The purpose of this study was to select features that will be used and overcome class imbalances that occur, so that the performance of the C4.5 algorithm worked more optimal in the classification process. The methods used as feature selection are PSO and bagging to overcome class imbalances. Classification was tested using the confusion matrix to determine the accuracy that was generated. From the results of this study, the application of PSO as a feature selection and bagging to overcome class imbalances with the C4.5 algorithm succeeded in increasing accuracy by 5.11% with an initial accuracy of 93.43% to 98.54%.
Copyrights © 2020