Natural Language Processing in Healthcare: Intelligent System for Extracting Insights from Clinical Text data

The World Health Organization identified COVID-19 to be a new pandemic in March 2020. This fatal virus spread over the world, affecting many different countries. Massive volumes of useful data were provided by social media platforms like Twitter on the outbreak to help evaluate health-related decision-making. To predict disease occurrence and issue early warnings, thus implement that user’s attitudes be evaluated through the use of efficient deep learning. The gathered tweets were pre-processed and divided into three categories like neutral, positive, and negative. In the second step, various features were extracted from the articles using a number of popular techniques to capture feature dataset, including TF-IDF, FastText, and Glove. The implemented TFIDF has combination of semantic features (Glove and FAsText) to accurately describe posts and enhance the classification process utilizing the deep learning method. The implemented TF-IDF-CNN method’s performance is evaluated using Twitter dataset. Implemented method achieved better performance than the existing intrusion detection methods including NLP-BiLSTM, TF-IDF, SVM and CF and the metrics are 96% of accuracy, 0.89 of f-measure, 0.93 of precision, and 0.90 of recall.