<![CDATA[This study aims to analyze the application of the Backward Elimination method in optimizing the Decision Tree C4.5 algorithm in predicting employee performance. The main problem in this study is the employee performance evaluation process which is still done manually so that it has the potential to cause inconsistency in the assessment results. The study used a dataset of 500 employee data with several performance assessment attributes such as age, education level, work discipline, productivity, and superior assessment. The research method includes data preprocessing, feature selection using Backward Elimination, application of the Decision Tree C4.5 algorithm, and model evaluation using 10-Fold Cross Validation in the RapidMiner application. The test results show that the Decision Tree C4.5 algorithm without optimization obtained an accuracy value of 89.60% and an AUC of 0.944. After applying the Backward Elimination method, model performance increased with an accuracy value of 92.80% and an AUC of 0.972. This increase indicates that the Backward Elimination method is able to reduce less relevant attributes so that the classification process becomes more optimal. Thus, the application of the Backward Elimination method has proven effective in improving the performance of the Decision Tree C4.5 algorithm in predicting employee performance.]]>
Copyrights © 2026
