Text

IMPLEMENTASI ARSITEKTUR CNN, BILSTM DAN ATTENTION BLOCK PADA KLASIFIKASI BERITA HOAKS DENGAN PENAMBAHAN AUGMENTASI BACK TRANSLATION DAN TEXTATTACK

---

Penilaian

---

The spread of hoax news has a negative impact on society and requires early detection of hoax news through classification. This research combines augmentation and classification methods. CNN is used to understand the relationship between adjacent words in part of the sentence, while BiLSTM understands the word order in the sentence as a whole, and attention block understands the global context simultaneously. This research requires large data through back translation and TextAttack techniques. Model performance evaluation is done by measuring accuracy, precision, recall, and f1-score values. The use of augmentation for hoax news data increases by 19.87%. Accuracy of 98.18% predicts most hoax news data correctly. Precision of 98.16% shows excellent accuracy in predicting both hoax and valid classes. Precision in the hoax class is higher than the valid class which is 98.45%, meaning that the model is right in predicting the hoax class. Recall 98.17% shows that it is sensitive to both hoax and valid classes. Recall of hoax class is higher than valid class which is 98.28% means the model is more sensitive to hoax class. F1-score of 98.46% shows a very good balance between precision and recall. The high f1-score value shows the consistency of the model in distinguishing hoax and valid classes. Based on the augmentation results, the number of data has increased from 20,292 to 24,324 data and the evaluation results of the combination of CNN architecture, BiLSTM and attention block can be categorized as very good in classifying hoax news in two classes.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2507001897	T169753	T1697532025	Central Library (Reference)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1697532025

Publisher

: Prodi Ilmu Matematika, Fakultas Matematika Dan Ilmu Pengetahuan Alam Universitas Sriwijaya., 2025

Collation

vii, 72 hlm.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

510.07

Content Type

-

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Matematika

Specific Detail Info

-

Statement of Responsibility

EM

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment