e-Health Education Using Automatic Question Generation-Based Natural Language (Case Study: Respiratory Tract Infection)

Abstract

In the midst of the outbreak, the public is flooded with information that is not necessarily true where hoax messages and fear spread faster than valid information and positive messages. For this reason, it is necessary to have consultation facilities that are accurate, fast and on target. This research creates an educational e-health in the form of a health question and answer system (with an upper respiratory tract infection case study) which can be used to provide answers that are more focused on valid information. Validation of information with answers searched using dynamic neural networks. Documents containing information are extracted to detect the correctness, and document keywords are associated with inquiries from users. The process of clarifying information uses automation of dynamic neural networks which allows the answers given to users to be more focused. The final result of this research is a virtual assistant that provides a corpus in the form of a QA-pair that can be generated automatically to provide accurate information to users working with upper respiratory tract infection with an accuracy of 71.6%.