Prediction modeling of annual parasite incidence (API) of Malaria in Indonesia using Robust regression of M-estimation and S-estimation

摘要

Malaria is an infectious disease caused by Plasmodium that lives and reproduces in human red blood cells. There is currently a large number of malaria cases in Indonesia. Several factors influencing the development of malaria in Indonesia is the ratio of the center of public healthcare, the percentage of households that have access to proper sanitation, the percentage of households that occupy livable homes, the number of general practitioners, and the number of people living in poverty. The relationship between malaria and its factors can be modeled using regression analysis. Regression analysis is usually estimated by using Ordinary Least Square (OLS) estimation. However, because there are outliers in the malaria dataset, it can be analyzed using robust regression. There are several estimation methods in robust regression, including M and S estimation. M estimation is an extension of Maximum Likelihood Estimation (MLE). It is robust, unbiased, and has minimum variance estimation. Meanwhile, S estimation is a family of high breakdown points, which is a general measure of the proportion of outliers data that can be handled before the observation affects the prediction model. The results show that the best model is a robust regression with S-Estimation. The factors that have a significant influence on the number of malaria in Indonesia are the percentage of cleaning water, the percentage of livable homes, and the number of the center public of healthcare in Indonesia with R~2 = 98.5%.