Natural Language Processing and Deep Learning Techniques for Fake Medical News Detection in Digital Healthcare Platforms

Abstract

The proliferation of popular digital information and communication platforms, and online information sharing of medical information, have greatly contributed to the dissemination of fake medical information, posing threats to public health, clinical decision making and medical awareness. The linguistic complexity, semantic ambiguity, and the vast amount of medical-related text data created in digital environments have made it difficult to detect misleading medical content. This research presents a Natural Language Processing (NLP) and deep learning-based approach for the detection of fake medical news in digital health care systems. The proposed framework comprises three key stages: text preprocessing, feature extraction, and intelligent classification, which aim to effectively distinguish fake and genuine medical news articles with high accuracy. Experimental evaluation and performance analysis is done using a fake news dataset for healthcare topics that consists of fake and factual medical information. Various deep learning models like Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (BiLSTM), and Bidirectional Encoder Representations from Transformers (BERT) are applied and compared. The experimental results show that the transformer-based models achieve good classification results, demonstrating the effectiveness of transformer-based models in reliable verification of digital healthcare content and detection of misinformation in healthcare.