Pelican Optimization-Assisted Deep Learning Framework for Adverse Drug Reaction Detection from Twitter Data

Adverse Drug Reactions (ADRs) lead to significant patient safety hurdles, necessitating timely and accurate detection to support pharmacovigilance efforts. Traditional ADR reporting systems suffer from underreporting and delays, prompting the need for alternative data sources such as social media. However, extracting meaningful insights from unstructured and noisy social media text presents substantial challenges. This research introduces a novel Deep Convolutional Recurrent Semantic Similarity Model (DCR-SSM), which integrates convolutional and recurrent layers with a semantic similarity mechanism and attention module to enhance ADR detection from Twitter data. The framework incorporates a robust preprocessing pipeline tailored to social media text and a novel Decision Tree-based Pelican Optimization Algorithm (DT-POA) for feature selection and Bag-of-Words encoding to capture relevant linguistic and semantic features. Simulation of the proposed framework is assessed on the SMM4H dataset, indicating the superior performance of the proposed Model over state-of-the-art ADR detection methods. The DCR-SSM accomplished an accuracy of 75%, 72%, recall of 72%, and an F1-score of 73%, outperforming traditional machine learning (SVM) and deep learning models (LSTM, Bi-LSTM, CNN). Compared to best-performing existing models, the proposed framework improves precision by up to 5.2% and maintains a balanced trade-off between recall and F1-score, ensuring better generalization in real-world applications. These findings highlight the potential of leveraging NLP and deep learning for mining patient-reported ADRs from social media, offering a scalable and cost-effective alternative to conventional pharmacovigilance methods. Future research can further explore multi-lingual ADR detection and domain-specific embeddings to enhance detection accuracy and adaptability across diverse healthcare settings.