Date of Award

Fall 2019

Date Available to Non-EWU Users

2020-12-12

Document Type

Thesis

Degree Name

Master of Science (MS) in Computer Science

Department

Computer Science

Abstract

Influenza has been identified by the World Health Organization as a global issue that could be more effectively served through an accelerated and widely-accessible public health surveillance tracking process. The ability to map and predict influenza outbreaks in a real-time heat map would be invaluable to health care systems to prepare for influenza outbreaks. In this study, the Twitter stream data is filtered to identify potential influenza-like illness (ILI) cases. Then the tracking of real-time influenza cases is further explored and analyzed through various machine-learning models. Among seven learning models developed to identify ILI tweets, the ELMo deep neural network model outperforms others regarding model accuracy and F-score. A heat map is generated to visualize real-time outbreaks of ILI in the U.S.A.

Share

COinS