Innovation in Automated Essay Scoring for Writing Assessment: A Case Study at SMA Negeri 1 Genteng

Andhini, Lestari
Maulana, Reza
Pratama, Naufal Rizky

Journal of Innovation in Education and Learning
Vol. 1 No. 3 (2025): OCTOBER 2025

This study aims to developt evaluate a Bidirectional Long Short-Term Memory (Bi-LSTM) based Automated Essay Scoring (AES) system to enhance the efficiency and objectivity of student writing assessment. A case study was conducted at SMA Negeri 1 Genteng using student essays graded by teachers as the ground truth. Evaluation results show QWK 0.83, MAE 0.41, and PCC 0.86, indicating a high agreement between model and manual scores. Beyond quantitative results, teachers and students perceived AES as efficient, objective, and supportive of writing improvement. All teachers (100%) agreed that AES speeds up grading, and 90% of students stated that automatic feedback helps them revise their writing. However, creativity still requires human judgment. Overall, the Bi-LSTM based AES system is feasible as an intelligent assessment tool in secondary education.