A Text Mining Approach to Discovering COVID-19 Relevant Factors

This paper describes a text mining approach that utilises the PyLucene search engine and the GrapeNLP grammar engine for extracting links between temperature, humidity and the spread of COVID-19, from a vast collection of scientific publications. The approach was developed in response to a Kaggle challenge from a consortium of research groups to develop text and data mining techniques that can assist the medical community in finding answers to a series of important questions on COVID-19. For this challenge, a large corpus of scientific publications known as the COVID-19 Open Research Dataset (CORD-19) was provided by the consortium. The approach presented in this paper was winner of the competition task of extracting key insights and building summary tables of COVID-19 relevant factors such as temperature and humidity.