Ikhsan, M (2021) EKSTRASI KATA KUNCI PADA RINGKARSAN MAKALAH ILMIAH MENGGUNAKAN BERT. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
07111850050001-Master_Thesis.pdf - Accepted Version Download (1MB) | Preview |
Abstract
Kata kunci berfungsi untuk memudahkan pencarian topik Makalah agar sesuai dengan keinginan pencari, penentuan kata kunci yang kurang tepat dapat mengakibatkan pencarian Makalah tidak sesuai dengan judul yang dimaksud. Untuk menanggulangi hal tersebut maka peneliti mengusulkan sebuah sistem pencarian kata kunci dalam abstrak Makalah menggunakan metode deep learning dengan model Bidirectional Encoder Representations from Transformers (BERT). Penelitian ini menggunakan 444 sampel data abstrak, pada proses pre-training peneliti mengunakan Natural Language Processing (NLP) untuk menghasilkan data yang lebih terstruktur dan tokinezer sebagai representasi input yang dapat diterima oleh BERT, selanjutnya peneliti melakukan training dengan model BERT. Hasil dari taining dilakukan pengujian akurasi dari standar pengukuran Main Average Precision (MAP) dengan nilai sebesar 55.1%.
================================================================================================
Keywords serve to facilitate the search for journal topics to match the wishes of the searcher, the determination of inappropriate keywords can result in journal searches not in accordance with the intended title. To overcome this, the researcher proposes a keyword search system in journal abstract using a deep learning method with a Bidirectional Encoder Representations from Transformers (BERT) model. This study uses 444 samples of abstract data, in the pre-training process the researcher uses NLP (Natural Language Processing) to produce more structured data and a Tokinezer as input representation that can be accepted by BERT, then the researcher conducts training with the BERT model. The results of the Taining test the accuracy of the Main Average Precision (MAP) measurement standard with a value of 55.1%.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Kata kunci, Deep learning, NLP, BERT, Tokenizer, MAP, Keywords, Deep learning, NLP, BERT, Tokenizer, MAP |
| Subjects: | T Technology > T Technology (General) > T57.74 Linear programming T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis |
| Depositing User: | M. Ikhsan |
| Date Deposited: | 14 Aug 2021 00:10 |
| Last Modified: | 17 Oct 2024 08:57 |
| URI: | http://repository.its.ac.id/id/eprint/86487 |
Actions (login required)
 |
View Item |