Adrian Kusuma Widjaja Kardana, - and Rizki Hikmawan, - (2025) PERANCANGAN MODEL MACHINE LEARNING UNTUK MENDETEKSI BERITA HOAKS MENGGUNAKAN ALGORITMA KLASIFIKASI. S1 thesis, Universitas Pendidikan Indonesia.
Abstract
Penelitian ini bertujuan mengembangkan aplikasi deteksi berita hoaks berbasis Natural Language Processing (NLP) dan algoritma Machine Learning dengan metode Research and Development (R&D) 4D, di mana tahap pengembangan aplikasi menggunakan model Waterfall. Dataset penelitian terdiri dari total 26.557 berita, yakni 14.665 berita valid dan 11.892 berita palsu. Empat model diuji: Logistic Regression, Naive Bayes, Random Forest, dan BERT. Berdasarkan evaluasi, Logistic Regression menghasilkan kinerja terbaik (accuracy 92,72%, precision 0,93, recall 0,93, F1-score 0,93) dengan performa stabil tanpa indikasi overfitting. Uji coba lapangan melibatkan 35 mahasiswa sebagai responden melalui pemberian materi, penggunaan aplikasi, dan post-test 10 soal yang telah divalidasi. Hasil uji One-Sample Wilcoxon Signed Rank Test menunjukkan p<0,001, menandakan aplikasi berpengaruh signifikan terhadap peningkatan literasi digital peserta. Temuan ini menunjukkan bahwa sistem tidak hanya akurat mendeteksi berita hoaks, tetapi juga efektif sebagai sarana edukasi publik. ----- This study aims to develop a hoax news detection application based on Natural Language Processing (NLP) and Machine Learning algorithms using the Research and Development (R&D) 4D method, with the development stage implemented through the Waterfall model. The dataset comprises 26,557 news articles, consisting of 14,665 valid and 11,892 hoax articles. Four models were tested: Logistic Regression, Naive Bayes, Random Forest, and BERT. Logistic Regression achieved the best performance (accuracy 92.72%, precision 0.93, recall 0.93, F1-score 0.93) with stable results and no significant overfitting. A field experiment involved 35 university students as respondents, including material delivery, application usage, and a 10-item validated post-test. The One-Sample Wilcoxon Signed Rank Test yielded p<0.001, indicating a significant effect of the application on improving participants’ digital literacy. These findings suggest that the system is not only accurate in detecting hoax news but also effective as a public educational tool
![[thumbnail of S_PSTI_2103456_Title.pdf]](http://repository.upi.edu/style/images/fileicons/text.png "S_PSTI_2103456_Title.pdf") |
Text
S_PSTI_2103456_Title.pdf Download (445kB) |
|
Text
S_PSTI_2103456_Chapter1.pdf Download (395kB) |
|
Text
S_PSTI_2103456_Chapter2.pdf Restricted to Staf Perpustakaan Download (525kB) | Request a copy |
|
Text
S_PSTI_2103456_Chapter3.pdf Download (504kB) |
|
Text
S_PSTI_2103456_Chapter4.pdf Restricted to Staf Perpustakaan Download (2MB) | Request a copy |
|
Text
S_PSTI_2103456_Chapter5.pdf Download (216kB) |
|
Text
S_PSTI_2103456_Appendix.pdf Restricted to Staf Perpustakaan Download (9MB) | Request a copy |
| Item Type: | Thesis (S1) |
|---|---|
| Additional Information: | https://scholar.google.com/citations?hl=en&user=EooADEoAAAAJ&scilu=&scisig=ACUpqDcAAAAAaK7255kk18Z93KZ_7JPUbFeQP4A&gmla=AH8HC4y6-q8aUgQp979kTzwcvFI6i71p-KaXHvRPgECep8FWmzQEcznQRTF1Oxn3UzU9aH6m62NrDgUEtEt0ksop4cEsu8pVI6B0eLk&sciund=8172611750231447171 ID SINTA Dosen Rizki Hikmawan : 6122897 |
| Uncontrolled Keywords: | Deteksi Hoaks, Natural Language Processing, Machine Learning Hoax Detection, Natural Language Processing, Machine Learning |
| Subjects: | L Education > L Education (General) T Technology > T Technology (General) |
| Divisions: | UPI Kampus Purwakarta > S1 Pendidikan Sistem Teknologi dan Informasi |
| Depositing User: | Adrian Kusuma Widjaja Kardana |
| Date Deposited: | 28 Aug 2025 03:57 |
| Last Modified: | 28 Aug 2025 03:57 |
| URI: | http://repository.upi.edu/id/eprint/136636 |
Actions (login required)
 |
View Item |