SISTEM QUESTION ANSWERING BAHASA INDONESIA UNTUK PERTANYAAN NON-FACTOID
Keywords:
non-factoid questions, pertanyaan non-factoid, question answering
Abstract
Fokus dari penelitian ini adalah untuk mengembangkan data dan sistem Question Answering (QA) Bahasa Indonesia untuk pertanyaan non-factoid. Penelitian ini merupakan penelitian QA non-factoid pertama untuk Bahasa Indonesia. Adapun sistem QA terdiri atas 3 komponen yaitu penganalisis pertanyaan, pengambil paragraf, dan pencari jawaban. Dalam komponen penganalisis pertanyaan, dengan asumsi bahwa pertanyaan yang diajukan merupakan pertanyaan sederhana, digunakan sistem yang berbasis aturan sederhana dengan mengandalkan kata pertanyaan yang digunakan (“apaâ€, “mengapaâ€, dan “bagaimanaâ€). Paragraf diperoleh dengan menggunakan pencarian kata kunci baik dengan menggunakan stemming ataupun tidak. Untuk pencari jawaban, jawaban diperoleh dengan menggunakan pola kata-kata khusus yang ditetapkan sebelumnya untuk setiap jenis pertanyaan. Dalam komponen pencari jawaban ini, diperoleh kesimpulan bahwa penggunaan kata kunci non-stemmed bersamaan dengan kata kunci hasil stemming memberikan nilai akurasi jawaban yang lebih baik, jika dibandingkan dengan penggunaan kata kunci non-stemmed saja atau kata kunci stem saja. Dengan menggunakan 90 pertanyaan yang dikumpulkan dari 10 orang Indonesia dan 61 dokumen sumber, diperoleh nilai MRR 0.7689, 0.5925, dan 0.5704 untuk tipe pertanyaan definisi, alasan, dan metode secara berurutan. Focus of this research is to develop QA data and system in Bahasa Indonesia for non-factoid questions. This research is the first non-factoid QA for Bahasa Indonesia. QA system consists of three components: question analyzer, paragraph taker, and answer seeker. In the component of question analyzer, by assuming that the question posed is a simple question, we used a simple rule-based system by relying on the question word used (“whatâ€, “whyâ€, and “howâ€). On the components of paragraph taker, the paragraph is obtained by using keyword, either by using stemming or not. For answer seeker, the answers obtained by using specific word patterns that previously defined for each type of question. In the component of answer seeker, the conclusion is the use of non-stemmed keywords in conjunction with the keyword stemming results give a better answer accuracy compared to non-use of the keyword or keywords are stemmed stem only. By using 90 questions, we collected from 10 people of Indonesia and the 61 source documents, obtained MRR values 0.7689, 0.5925, and 0.5704 for type definition question, reason, and methods respectively.
Published
2012-05-29
How to Cite
Purwarianti, A., & Yusliani, N. (2012). SISTEM QUESTION ANSWERING BAHASA INDONESIA UNTUK PERTANYAAN NON-FACTOID. Jurnal Ilmu Komputer Dan Informasi, 4(1), 10-14. https://doi.org/10.21609/jiki.v4i1.151
Section
Articles
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
