COVID-19-Related Predictions Using NER on News Headlines

Today, all the newspapers and online news content are flooded with the news of coronavirus (i.e., COVID-19). The virus has spread across the globe at an alarming rate. Thus, people need to remain updated about news regarding the ongoing pandemic which has taken whole world by storm. Therefore, named entity recognition (NER) is applied to extract important information from these news headlines and articles and further used for more applications related to COVID-19 in India. This chapter uses the SpaCy module to categorize the tokens extracted from the news headlines database into various pre-defined tags. Further, four different machine learning models, namely CRF Model, LSTM Model, LightGBM Model, and AdaBoost Model, are applied for performing tagging. After that, these tags are used to predict different information regarding COVID-19. Some of these applications include finding nearby hospitals and pharmacies, predicting future potential hotspots in India, worst affected states of India, gender-based comparisons, age group-based comparisons, and area-based spreading of the virus.