<![CDATA[ABSTRACTThis study aims to classify the success factors of job training participants at LPK Ziona Gunungsitoli by comparing Support Vector Machine (SVM) and Decision Tree C4.5. The dataset used in this study was obtained from participant records for 2022-2025 and contains tracer study outcomes, gender, education level, training type, program type, and year of graduation. Personal identifiers such as names, national identity numbers, addresses, and telephone numbers were excluded during preprocessing. The tracer study variable was transformed into a binary target: participants who were working or continuing study were categorized as successful, while participants who were still looking for work were categorized as not yet successful. From 376 raw records, 372 valid records were processed. The data were encoded using one-hot encoding and evaluated using an 80:20 stratified train-test split. The results show that Decision Tree C4.5 achieved an accuracy of 81.33%, precision of 94.83%, recall of 83.33%, and F1-score of 88.71%. Meanwhile, SVM achieved an accuracy of 77.33%, precision of 94.55%, recall of 78.79%, and F1-score of 85.95%. The most influential attributes in the Decision Tree model were year, training type, gender, education, and program type. These results indicate that Decision Tree C4.5 is more suitable for this dataset because it provides both higher performance and interpretable rules for institutional decision support.]]>
Copyrights © 2026
