<![CDATA[Vocational high school students are prepared to enter the labor market. However, field facts show that many graduates work outside their field of expertise. To address this, the special job placement bureau (Bursa Kerja Khusus/BKK) plays an important role as it connects graduates with industry. In addition, BKK provides pre-employment training such as interview preparation and soft-skill development. This study aims to develop a classification-based career-path mapping system integrated with BKK functions. The data used are scores of eight competency dimensions for vocational students obtained from BKK. The method employed is the Random Forest algorithm. We conduct hyperparameter tuning with cross-validation. Results show Random Forest achieves accuracy of 0.895 and an F1 score of 0.905. These results indicate that optimizing for F1 yields the best balance between precision and recall while maintaining high overall accuracy. Overall, this study confirms a trade-off between overall accuracy and inter-class balance (F1): constrained tree depth tends to maximize accuracy, whereas unconstrained depth benefits F1. Random Forest proves reliable and stable for the classification task in this career-path mapping]]>
Copyrights © 2025
