Bio-Inspired Artificial Intelligence with Natural Language Processing Based on Deceptive Content Detection in Social...

Area of Science:

• Artificial Intelligence
• Natural Language Processing
• Machine Learning

Background:

• Fake news detection on social media is a growing concern.
• Existing Machine Learning (ML) and Deep Learning (DL) models are widely used.
• There is a need for improved techniques to identify deceptive content.

Purpose of the Study:

• To propose a novel Bio-inspired Artificial Intelligence with Natural Language Processing Deceptive Content Detection (BAINLP-DCD) technique.
• To enhance the accuracy of fake news detection in social networking environments.

Main Methods:

• The BAINLP-DCD technique employs data preprocessing for data transformation.
• A Multi-Head Self-attention Bi-directional Long Short-Term Memory (MHS-BiLSTM) model is utilized for deceptive content detection.
• The African Vulture Optimization Algorithm (AVOA) optimizes the hyperparameters of the MHS-BiLSTM model.

Main Results:

• The BAINLP-DCD technique demonstrated enhanced performance in detecting fake news.
• Maximum accuracy of 92.19% was achieved on the BuzzFeed dataset.
• Maximum accuracy of 92.56% was achieved on the PolitiFact dataset.

Conclusions:

• The proposed BAINLP-DCD technique effectively detects deceptive content on social media.
• The integration of bio-inspired algorithms and DL models shows promise for fake news detection.
• The study highlights the potential of BAINLP-DCD for improving social media information integrity.