Predicting Crimean-Congo Hemorrhagic Fever Outbreaks via Multivariate Time-Series Classification of Climate Data

Abstract

Crimean-Congo hemorrhagic fever (CCHF) is a vector-borne disease that is spread by ticks (specifically of the Hyalomma marginatum species) and is influenced by climate patterns. CCHF has a fatality rate ranging from 3-50% for humans and is a high-priority disease among international health organizations. We hypothesize that temporal variability in climate variables (temperature and precipitation) can be used to predict CCHF outbreaks in a particular region. There is a need to analyze the effects of climatic patterns on the spread of CCHF to allow high-risk countries to better prepare for possible outbreaks. We propose an approach that utilizes multivariate time-series classification (MTSC) to detect temporal climatic patterns and predicts reports of CCHF outbreaks within Pakistan with a 91.5% test accuracy.