<![CDATA[Manual essay grading at the vocational school level is a time-consuming and subjective process. This research implemented and evaluated an automatic essay scoring model using a combination of the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm for word weighting and Word Embedding for semantic meaning analysis. The model was tested using a dataset of 360 essay answers from 36 students at SMK Budi Bakti Ciwidey with train-test split validation. The tuning process on the training data showed that a weighting that prioritized semantic analysis (90% Word Embedding) provided the best performance. In the final testing on 90 test data, the model achieved an excellent Mean Absolute Error (MAE) of 6.80, but with a weak Pearson correlation of 0.12 against the teacher's scores. This research concludes that the proposed model is successful in generating scores that are very close to the teacher's scores (low MAE), but still has limitations in terms of scoring consistency (weak correlation), which is influenced by the quality of the key answers and an imbalanced dataset.]]>
Copyrights © 2026
