A Natural language processing model to detect outbreaks of infectious diseases in Kenya

Abstract

Infectious diseases are more prevalent in urban settings than in rural areas. This is because the population in the metropolitan areas is higher, and several resources like public transport and social activities are shared despite the large population. The government is usually slow to react whenever there is an outbreak due to many reasons. Outbreaks of contagious diseases exert considerable pressure on the public and hospital facilities. Kenya is usually affected by epidemics of different kinds and at various times. The government gets information about outbreaks mainly through the health sector when several cases are reported in a specific region. Social media platforms enable users to share information at a global level. Data collected from these platforms can be utilised to monitor and predict disease outbreaks. X, one of the largest social media platforms in the world, hosts a vast amount of user-generated content, most of which is publicly accessible. Tweets generated by Kenyan citizens will be analysed for this study, and this study explores the application of Natural Language Processing (NLP) models to detect and predict disease outbreaks in Kenya by analysing publicly available data sources such as social media. This research seeks to help the government and other stakeholders know when there is an outbreak of infectious diseases and to put up stringent measures to mitigate further spread. This study delves into the development of a natural language processing model that will extract data from social media, e.g. X. The model focuses on information related to a particular infectious disease, and it utilises machine learning techniques to process unstructured textual data, identifying patterns and signals indicative of emerging public health threats. By leveraging real-time data, the NLP model aims to provide early warnings of outbreaks, enabling quicker responses from public health authorities. The results demonstrate the model's potential in enhancing disease surveillance. Key words: Infectious diseases, infodemic, public health surveillance, social media, syndromic surveillance