An Analysis of Methods for Forecasting Epidemic Disease Outbreaks using Information from Social Media

Abstract

Predicting the rise or fall of an epidemic or pandemic is an essential part of establishing control over it. Post-World War 1, when there was an outbreak of the “Black Plague” there weren’t any means to analyze and predict. Although today we are equipped with tools like Machine Learning and Artificial Intelligence which have certainly enabled us to prevent unnecessary loss of life. It helps prepare the health officials to build the infrastructure and interpret the intensity of preparedness regulation of resources. The aim of this survey is to analyze and shed some light on the various algorithms and methods such as - regression models, neural networks, ARIMA, etc. Before building any model, gathering and processing the data is also essential. Hence our paper also focuses on which social media platforms proved beneficial in comparison to all we found and then made fit to be incorporated into the models. While researching for this paper, we observed that every disease has a different transmission type that leads to an outbreak and is a key factor in constructing a model. The literature evaluation in this work is centered on various prediction algorithms and their strategies for extracting online data from social media sites like Facebook and Twitter, all of which have drawn a lot of interest in early disease diagnosis for public health.