EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

Multimodal Hoax News Detection Using OCR and a Multi-Kernel 1D-CNN Model

Seggy Ferdianza Agustin

Journal of Artificial Intelligence and Data Science · 2025

DOI: 10.67561/jaids.v1i2.3 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Artificial Intelligence and Data Science, 2025. Verfasst von Seggy Ferdianza Agustin. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The focus of this research is to develop an autonomous Indonesian hoax news detection system that can leverage textual information from both news articles and text within images or videos. The proposed approach combines Optical Character Recognition (OCR) for extracting text from visual data and a Multi-Kernel One-Dimensional Convolutional Neural Network (Multi-Kernel 1D-CNN) for modeling linguistic patterns across multiple n-gram scales. The dataset is compiled from two credible Indonesian sources: Kompas.com for accurate information and TurnBackHoax.Id to hoaxes that we have verified, which sum up to 24439 labeled samples in a nearly balanced distribution of classes (50.89% valid news and 49.11% hoax news). Model performance is assessed using an 80:20 hold-out split and 10-fold stratified cross-validation. Experiments show that the proposed Multi-Kernel 1D-CNN exhibits powerful performance, but with an average accuracy of 99.918% ± 0.055, precision of 99.950% ± 0.058, recall of 99.883% ± 0.112, and F1-score of 99.917% ±0.056, which consistently outperforms the single-kernel CNN baseline method in different temporal-length test samples we have discussed above. In conclusion, the OCR and Precision–Recall curves also suggest almost perfect separability between the two classes. In summary, OCR with a Multi-Kernel 1D-CNN is an effective and efficient multimodal approach for detecting hoax news. It can be used in real-time decision-support systems for Indonesian online news.

Zitieren

APA

Agustin, S. F. (2025). Multimodal Hoax News Detection Using OCR and a Multi-Kernel 1D-CNN Model. Journal of Artificial Intelligence and Data Science. https://doi.org/10.67561/jaids.v1i2.3

Vancouver

Agustin SF. Multimodal Hoax News Detection Using OCR and a Multi-Kernel 1D-CNN Model. Journal of Artificial Intelligence and Data Science. 2025. doi:10.67561/jaids.v1i2.3

BibTeX
@article{Agustin2025,
  author = {Agustin, Seggy Ferdianza},
  title = {Multimodal Hoax News Detection Using OCR and a Multi-Kernel 1D-CNN Model},
  journal = {Journal of Artificial Intelligence and Data Science},
  year = {2025},
  doi = {10.67561/jaids.v1i2.3}
}
Kostenlos starten →