In previous research, machine learning methods have been carried out in implementing automatic essay assessment. Automated essay scoring (AES) is used in evaluating and assessing student essays written based on the questions given. However, there are difficulties in carrying out an automatic assessment carried out by the system, the difficulty occurs because of typing errors, use of regional languages, and incorrect punctuation. These errors make the assessment less consistent and accurate. One of the problems in essay answers is the proofing process which is more complicated than multiple-choice questions. The Automated Essay Scoring established to manage the process of essay scoring assessment and evaluation using computation approach which is machine learning with classification capabilities. The problem occurs when we have unbalanced dataset and a few labeled data in especially for the model training process. The evaluation conducted with comparing the classification model with several combinations. This study proposes to analyze and comprehensive evaluation of the Word Embedding with GRU and RNN, TFIDF with AdaBoost, back of word with AdaBoost, and FastText with MLP which are expected to solve these two problems. The optimal model and architecture for sequential feature-based scoring (RNN) with a fairly stable performance.
Rumpun Ilmu
Ilmu Komputer
Bahasa Asli/Original Language
English
Level
Internasional
Status
Dokumen Karya
No
Judul
Tipe Dokumen
Aksi
1
Acceptance.pdf
Bukti Accepted
2
The Development of Automated Essay Scoring Based on Classification Models.pdf
