A Hybrid Approach for Integrating Deep Learning and Explainable AI for augmented Fake News Detection

Abstract

In today's digital age, the proliferation of misinformation poses significant challenges to the integrity of news consumption. This study aims to develop an augmentedIntegrated Hybrid Deep Learning AI (IHDLAI) framework, leveraging Deep learning and Artificial Intelligence (AI) enhanced natural language processing techniques to effectively identify and mitigate the spread of false information. The proposed framework comprises three distinct phases: Data Collection and Pre-processing, Model Development and Training, and Evaluation.During the Data Collection and Pre-processing phase, diverse datasets were meticulously curated from verified news outlets, social media, and fact-checking websites, ensuring a balanced representation of fake and genuine news. The collected data underwent extensive cleaning and pre-processing, including tokenization, normalization, and feature extraction, resulting in a robust dataset ready for modelling.In the Model Development and Training phase, various Deep Learning models, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs) with LSTM cells, and transformer-based models like BERT, were trained and optimized. The final phase, Evaluation and Deployment, involved a comprehensive assessment of the models on a test set to ensure their efficacy. The results of IHDLAI demonstrated a significant improvement in detecting fake news, with the best model achieving an accuracy of 93%, a precision of 92%, a recall of 91%, and an F1 score of 91.5%. The findings underscore the effectiveness of the proposed framework in combating misinformation, providing a reliable tool for enhancing the credibility of news consumption.