Hoax news is false and misleading information that can cause provocation and hatred for readers. With easy internet access, the spread of hoax news is getting more massive. Therefore, there needs to be a method that can detect hoax news. The research uses deep learning methods by integrating text mining to find information and news patterns related to hoaxes. By using a dataset from the kaggle site totaling around 2700 then text preprocessing is carried out so that the data is more structured for further processing. Then make feature engineering from BERT so that the data can be processed by machine learning with three classification methods namely BERT, SVM and random forest then testing and evaluation. In this study, the model that produces the highest performance is BERT with (accuracy = 0.99, ROC-AUC = 0.99) compared to traditional machine learning models.
Copyrights © 2024