| Abstract (EN) |
The objectives of this research were: (1) to investigate the effectiveness of the Long
Short-Term Memory (LSTM) technique for fake news detection, and (2) to compare the
performance of the LSTM technique with baseline machine learning methods in fake news
detection. The study was conducted using real-world fake news data collected from global
news sources, consisting of a total of 20,800 news articles, including 10,413 fake news articles
and 10,387 real news articles. The dataset was divided into a training set (80%) and a testing
set (20%). News classification was performed using a deep learning approach based on the
LSTM neural network. The results showed that: (1) the LSTM neural network achieved high
effectiveness in fake news detection, with an accuracy of 97.78%; and (2) when compared
with baseline techniques, namely Decision Tree and Random Forest, the LSTM model
demonstrated superior performance. This performance gain can be attributed to the deep
learning capability of the LSTM model in capturing complex sequential patterns in textual
data. |
Creative Commons License