Simatupang, Stephanie Hebrina Mabunbun (2026) Automatic Essay Scoring (AES) untuk Penilaian Tata Bahasa Pada Esai Bahasa Indonesia Dengan Metode Hybrid Linguistic Dan IndoBERT. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5027221069_Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5027221069_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
Penilaian esai secara manual oleh tenaga pengajar kerap menghadapi sejumlah tantangan, seperti tingginya kebutuhan waktu, risiko kelelahan, dan potensi inkonsistensi akibat subjektivitas penilai. Kondisi ini menjadi semakin kompleks seiring meningkatnya jumlah peserta didik dan volume tugas berbasis esai, sehingga mendorong kebutuhan akan sistem penilaian otomatis yang mampu memberikan evaluasi yang cepat, objektif, dan stabil. Untuk menjawab kebutuhan tersebut, penelitian ini mengembangkan Sistem Penilaian Esai Otomatis untuk teks berbahasa Indonesia dengan pendekatan model hybrid. Pendekatan ini mengombinasikan fitur linguistik meliputi aspek ejaan, tata bahasa, serta karakteristik statistik teks dengan representasi embedding kontekstual dari model pra-latih IndoBERT. Pengujian dilakukan melalui tiga skenario eksperimen menggunakan algoritma Support Vector Classifier (SVC), yaitu skenario berbasis fitur Linguistik, fitur IndoBERT, dan pendekatan hybrid. Hasil evaluasi menunjukkan bahwa pendekatan hybrid menghasilkan performa terbaik dengan nilai Quadratic Weighted Kappa (QWK) sebesar 0,902. Nilai ini melampaui performa skenario Linguistik 0,885 maupun skenario IndoBERT 0,841, sehingga menegaskan efektivitas integrasi kedua jenis fitur dalam meningkatkan akurasi penilaian. Sebagai pelengkap evaluasi model, sistem AES berbasis web dikembangkan menggunakan Streamlit untuk mendukung prediksi skor esai secara otomatis serta penyediaan feedback linguistik.
=============================================================================================================================
Manual essay scoring by educators frequently encounters several challenges, such as significant time demands, the risk of fatigue, and potential inconsistency resulting from grader subjectivity. This situation becomes increasingly complex with the rising number of students and the volume of essay-based assignments, thereby necessitating an automated scoring system capable of providing rapid, objective, and consistent evaluations. To address this need, this research develops an Automatic Essay Scoring (AES) system for Indonesian texts utilizing a Hybrid model approach. This approach combines linguistic features encompassing spelling, grammar, and statistical text characteristics with contextual embedding representations from the pre-trained IndoBERT model.Testing was conducted through three experimental scenarios using the Support Vector Classifier (SVC) algorithm, a Linguistic feature-based scenario, an IndoBERT feature-based scenario, and the Hybrid approach. Evaluation results indicate that the Hybrid approach yielded the best performance, achieving a Quadratic Weighted Kappa (QWK) score of 0,902. This score surpasses the performance of both the Linguistic scenario 0,885 and the IndoBERT scenario 0,841, thereby confirming the effectiveness of integrating these feature types in enhancing scoring accuracy. Furthermore, the developed AES system has been implemented as a web-based application, enabling users to upload documents to obtain automatic score predictions and receive immediate linguistic feedback.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Kata kunci: Automatic Essay Scoring, IndoBERT, Fitur Linguistik, QWK, Metode Hibrid. Keywords: Automatic Essay Scoring, IndoBERT, Linguistic Features, QWK, Hybrid Method. |
| Subjects: | T Technology > T Technology (General) > T57.5 Data Processing |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis |
| Depositing User: | Simatupang Stephanie Hebrina Mabunbun |
| Date Deposited: | 20 Jan 2026 06:19 |
| Last Modified: | 21 Jan 2026 07:02 |
| URI: | http://repository.its.ac.id/id/eprint/129825 |
Actions (login required)
 |
View Item |