<![CDATA[The mental readiness of student organization cadres, particularly those in the Muhammadiyah Student Association (IMM), has traditionally been evaluated qualitatively and subjectively, making data-driven decisions regarding workload distribution and cadre development difficult. This study aims to classify the mental readiness of IMM cadres at Universitas Muhammadiyah Kalimantan Timur (UMKT) as an initial basis for a data-driven organizational human-resource evaluation system. The study applies a multimodal Machine Learning approach by combining semantic representations of interview transcripts extracted using IndoBERT with nonverbal observation scores covering self-confidence, communication fluency, and narrative depth. Data were collected from 16 in-depth interviews and expanded to 32 samples using Random Deletion augmentation with a 20% probability. The 768-dimensional text representation was concatenated with three observation features to form a 771-dimensional multimodal feature vector and classified using XGBoost. Evaluation was conducted using Stratified 3-Fold Cross-Validation with scale_pos_weight to address class imbalance. The model achieved an overall accuracy of 88%, with precision and recall of 91% for the “Ready” class and 80% for the “Vulnerable” class. These results indicate that combining textual information with nonverbal observations provides a sufficiently strong classification signal for distinguishing cadre mental readiness. However, because only 16 original data samples were available, this study should be regarded as a proof of concept rather than a generalizable performance benchmark. Further validation using larger and more diverse samples is required to assess model generalization.]]>
Copyrights © 2026
