Skripsi

KLASIFIKASI TEKS LOWONGAN KERJA ASLI DAN PALSU MENGGUNAKAN BIDIRECTIONAL ENCODER REPRESANTATIONS FROM TRANSFORMERS (BERT)

---

Job vacancy information is now more accessible to the public. However, this ease of access also has a negative impact, namely the increased ease with which fake job vacancies can be spread, causing harm to job seekers. Therefore, this study aims to develop a model based on Bidirectional Encoder Representation from Transformers (BERT) to classify genuine and fake job vacancies. The dataset used is a secondary dataset with a total of 44,045 data points, consisting of 31,522 genuine labeled data and 12,523 fake labeled data. The labeled dataset then underwent preprocessing and fine-tuning to find the model with the best performance. The model was trained with six test scenarios using learning rate variations of 2e-5 and 3e-5 and a number of epochs of 3, 4, and 5. After fine-tuning with six different scenarios, the model from the third scenario with a learning rate of 2e-5 and 5 epochs produced the best performance with an accuracy value of 99.76%, precision of 99.73%, recall of 99.41%, and an F1-score of 99.57%.

---

Availability

Detail Information

Series Title

-

Call Number

T1897572025

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2025

Collation

xv, 141 hlm.; ilus.; tab.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.320 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Pemrosesan Bahasa Alami--Jaringan Syaraf

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment