Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal EMITTER International Journal of Engineering Technology
Akmal -
Padjadjaran University
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Consistency of News Titles and Contents With Multi Level Classification Using Bidirectional Encoder Representations from Transformer (BERT) Afrida Helen; Akmal -; Muhammad Razzaaq Fadilah
EMITTER International Journal of Engineering Technology Vol 14 No 1 (2026)
Publisher : Politeknik Elektronika Negeri Surabaya (PENS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24003/emitter.v14i1.1097

Abstract

News is written in a standard format that consist of a title followed by the content. The title must be present a content of news. Sometimes found that news title did not represent the content, causing readers to be disappointed when they finish reading the whole news. This research proposes multi-class and multi level classification using BERT algorithm to detect the suitability of news titles and their contents. The first level performs classification for complete data and the second level performs classification for data that has been separated into 2 parts, Title dan Conten. The dataset pertains to the realm of Covid-19 news because its substantial volume and diverse range of conversation topics. It is sufficient to be used as a dataset. We select several news topic classes that are often discussed, that are health, politic, and economy. In order to determine suitability, we categorise both the news Title and the news Content into predefined topic classes. The dataset exhibits an inbalance for each topic class. This research proposed a model to determine the appropriateness of both the news Title and the news Content. We use Bidirectional Encoder Representations from Transformers for building model. BERT is one of the state-of-the-art in the field of Natural Language Processing (NLP). Multi-layer classification that separates dataset (title and content) determine the consistency between them is very effective in increasing accuracy, and the BERT model is very supportive in solving linguistic problems well. This research succeeded in providing an accuracy 15% greater compared to previous research.
Co-Authors Afrida Helen Muhammad Razzaaq Fadilah