Related Experiment Video
Updated: Jun 28, 2026
03:31
End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
Published on: December 15, 2023
A hybrid deep learning framework for fake news detection using LSTM-CGPNN and metaheuristic optimization
Ramesh Kumar Ayyasamy1, Chinnasamy Ponnusamy2, Kovvuri N Bhargavi3
1Faculty of Information and Communication Technology, Universiti Tunku Abdul Rahman, Kampar, Perak, Malaysia. rameshkumar@utar.edu.my.
Scientific Reports
|November 25, 2025
Summary
This study introduces a hybrid deep learning model for effective fake news detection on social media. The novel approach significantly enhances accuracy and robustness in identifying deceptive content.
Area of Science:
- Artificial Intelligence
- Computer Science
- Social Media Analysis
Background:
- The pervasive spread of fake news on social media platforms poses significant challenges to public trust and informed decision-making.
- Traditional fake news detection methods often struggle with the complexity and volume of online information.
- There is a critical need for advanced, robust, and scalable solutions to identify and mitigate the impact of misinformation.
Purpose of the Study:
- To develop and evaluate a hybrid deep learning model for improved fake news detection.
- To enhance the accuracy and robustness of fake news identification by integrating multiple advanced techniques.
- To provide a scalable and efficient solution for real-world applications in combating misinformation.
Main Methods:
- A hybrid deep learning framework combining Long Short-Term Memory (LSTM) networks for feature extraction and Convolutional Gaussian Perceptron Neural Networks (CGPNN) for classification.
- Integration of the Moth-Flame Whale Optimization (MFWO) algorithm for optimizing model hyperparameters.
- Experimental validation on benchmark datasets (ISOT, Fakeddit, BuzzFeedNews, FakeNewsNet) using TF-IDF text representation and standardized preprocessing.
Main Results:
- The proposed hybrid model achieved high performance, with up to 98% accuracy and 95% F1-score.
- Statistically significant improvements (p < 0.05) were observed compared to existing transformer-based and graph neural network models.
- The model demonstrated superior performance, delivering 3-8% higher accuracy and F1-score than state-of-the-art approaches.
Conclusions:
- The hybrid deep learning framework effectively captures linguistic nuances and textual anomalies characteristic of fake news.
- The developed model offers a robust, scalable, and efficient solution for large-scale fake news detection.
- This research has practical implications for social media monitoring, digital journalism, and public awareness initiatives against misinformation.