Multimodal visual analysis of vector-borne infectious diseases

Abstract

In this work, we mainly analyze an infectious disease – dengue, which is transmitted by Aedes aegypti. Here, we propose a visual analysis method based on multiple perspectives. At first, visual analysis is used to calculate the probability of occurrence of new cases of dengue and the relative risk of occurrence of dengue cases influenced by various variables causing dengue. As a result, we find that climatic variables (rainfall, maximum, minimum and average temperature), imported case and density of Aedes aegypti are three major factors affecting outbreak of dengue. Then, prediction model is built based on analysis to realize early warning. At last, various visual methods are used to show prediction results. This whole method is first used in Yunnan province, China. Compared with current methods, our method takes more factors into consideration and bases on machine learning to build prediction model, which can improve the accuracy of prediction.