A Highly Accurate Adverse Drug Reactions (ADR) Detection from Medical Forum Comments Using Long Short-Term Memory Networks

Adverse drug reactions (ADRs) are only one example of the kind of helpful medical information that is frequently accessible on social media platforms for healthcare, where people can share their own experiences with treatments on desktop computers and mobile devices. A growing number of researchers are interested in gathering this valuable medical data from social media. This research explores the effects of three aspects on the recognition of ADR mentions in social media for the medical field and proposes a deep neural network of long short-term memory (LSTM) neural networks to do so. The texts from the dataset are pre-processed, and characteristics such as the semantic feature, text statics, and American Standard Code For Information Interchange (ASCII) array are extracted. Further, the features are converted into LSTM networks to perform the testing operation using the above features. This work was evaluated using two datasets, CODEC and ADR Corpus. Via extensive experiments, this work achieved 99.79 accuracies, 98.37 sensitivity, 97.63 specificities, 99.72 precision, and 98.39 recall 97.62 F1-score for the CODEC dataset. 98.16 for accuracy, 99.19 for sensitivity, 98.49 for specificity, 99.49 for precision, 96.72 for recall, and 93.16 for F1-score for ADE corpus, respectively.