| Autor: |
Yunita Sari, Indra Hidayatulloh, Isna Alfi Bustoni, Bambang Nurcahyo Prastowo, Guntur Budi Herwanto |
| Rok vydání: |
2018 |
| Předmět: |
|
| Zdroj: |
ICEAP Proceeding Book Vol 2. |
| DOI: |
10.26499/iceap.v2i1.95 |
| Popis: |
This paper presents UKARA, a fast and simple automatic short-answer scoring system for Bahasa Indonesia. Automatic short-answer scoring holds an important role in speeding up automatic assessment process. Although this area has been widely explored, only very limited number of previous work have studied Bahasa Indonesia. One of the major challenges in this field is the different type of questions which require different assessments. We are addressing this problem by implementing a combination of Natural Language Processing (NLP) and supervised machine learning techniques. Our system works by training a classifier model on human-labeled data. Using three different types of Programme for International Student Assessment (PISA) student responses, our system successfully produced the F1-score above 97% and 70% on dichotomous and polytomous scoring types respectively. Moreover, UKARA provides a user-friendly interface which is simple and easy to use. UKARA offers a flexibility for human grader to do re-scoring and re-training the model until the optimal performance is obtained. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
