Faculty Mentor(s)
Yew-Meng Koh, Mathematics
Document Type
Poster
Event Date
4-21-2017
Abstract
The Aedes genus of mosquito is the vector for at least three viral diseases - Dengue Fever (DF), Zika and Chikungunya. In many regions of the world, an upward trend in DF and Zika infections is observed. Using disease data from the Singaporean Ministry of Health and population data from the Singapore Department of Statistics, various statistical models are fit to DF, Dengue Hemorrhagic Fever (DHF; an often-fatal complication of Dengue Fever) and Chikungunya. These models differ in their consideration of the overall data structure, and have different underlying assumptions. Of particular interest is a prediction model based on neural networks, which we present. The merits and performance of these models are discussed and the accuracy of predictions made by each model are compared. The statistical method for determining prediction bounds for the neural network model is also discussed. These prediction models provide an objective method for public health management and policy making.
Recommended Citation
Repository citation: Sandgren, Matt, "Predicting Dengue Fever Incidence" (2017). 16th Annual Celebration of Undergraduate Research and Creative Performance (2017). Paper 103.
https://digitalcommons.hope.edu/curcp_16/103
April 21, 2017. Copyright © 2017 Hope College, Holland, Michigan.
