Analisis Sentimen Penerapan Deep Learning dan Analisis Sentimen terhadap Gap Kompetensi Lulusan Lembaga Pendidikan dan Pelatihan Vokasi terhadap Dunia Kerja dengan Metode Long Short-Term Memory (LSTM)

Yahya, Susilawati
Sitorus, Zulham
Iqbal, Muhammad
Nasution, Darmeli
Farta Wijaya, Rian

Bulletin of Information Technology (BIT)
Vol 6 No 2: Juni 2025

The gap between vocational graduates’ competencies and labor market demands remains a pressing issue in Indonesia. This study aims to analyze alumni perceptions regarding the alignment between competencies acquired during their studies at LP3I Banda Aceh and real-world job requirements. A quantitative approach was adopted using a deep learning method based on Long Short-Term Memory (LSTM). Data were collected through an online survey containing open-ended responses from 934 alumni, followed by preprocessing, tokenization, lexicon-based sentiment labeling, and data splitting into training and testing sets. The models developed included pure LSTM, LSTM with class weights, and Bidirectional LSTM (BiLSTM). Results indicate that BiLSTM achieved the highest performance with 90% accuracy and a weighted F1-score of 0.91. Additionally, 44.5% of respondents expressed neutral or negative sentiments, highlighting a mismatch between acquired competencies and industry demands. These findings underscore the urgency of curriculum evaluation and stronger collaboration between vocational institutions and the labor market. This study demonstrates that deep learning offers an efficient and objective tool for competency mapping in vocational education.