Open AccessExtraction of keywords in the abstract of scientifics papers using BERT
Author(s) -
Mohammad Robikhul Ikhsan,
Reza Fuad Rachmadi,
Eko Mulyanto Yuniarno
Publication year - 2021
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/1175/1/012016
Subject(s) - computer science , keyword extraction , information retrieval , keyword search , transformer , encoder , word (group theory) , artificial intelligence , natural language processing , mathematics , physics , geometry , quantum mechanics , voltage , operating system
Keywords are used to facilitate searches, inaccurate keyword determination, which can result in searches for journals that do not match the intended title. To overcome this, it is necessary to implement a keyword search in the abstract using deep learning methods from a journal to produce the right keywords that will distinguish journals from one another. The purpose of this study is to implement a keyword search system in abstracts using the Bidirectional Encoder Representations from Transformers (BERT) method as an abstract word search system that produces keyword answers that have context according to the desired keyword results. The results of this study get optimal results in the epoch 500, and from these results using a dataset with of 444 training data and 10 test data for testing.