Regression modeling of dengue incidence using time-lagged meteorological factors

Abstract

We develop simple models using three regression techniques — linear model LM, power law model PM, and generalized linear model GLM — to investigate the influence of explanatory variables C with responses variables d in the National Capital Region (NCR). We use the C with various time lags to determine the highest coefficient of determination r², which we employed to develop the models. Subsequently, we determined which model, utilizing C data from stations SG, N, and PA, best described the distribution of dengue incidences in the NCR based on their corresponding r² values. Our results show that linear model LM with the data from N station best explains the variability of dengue cases in the NCR.