Abstract:
Ever-increasing demand of quality health care services, service delivery institutions and service providers struggle to meet up excess demands, particularly associated with peak events like diarrhea. Front-line health delivery services and providers are not usually adequately pre-informed and do not have adequate resources to meet the needs of health care. The main motivation of this study is to model the infectious diarrheal patients hospitalized data considering its variations over time as the dynamics of the disease depends on various climate factors. The monthly data were used the number of infected diarrheal patients who were hospitalized for diarrhea from January, 1993 to December, 2017 as target and IOD & MEI as explanatory.
The selected models were those with appropriate distribution for count data, flexible and that allows insertion of explanatory variables experimented. In literature, it is found that Poisson and Negative Binomial distributions have been widely used in GLM with recurrence, but its performance is weak in comparison to the Wavelet-INGARCH model. This research finds suitable forecasting models for time series counts of infectious diarrheal patients, and demonstrates expected results for decision-maker related to diarrheal disease. The practical consequence behind the research was to incorporate wavelet decomposition whether the data are either stationary or non-stationary or linear or non-linear or seasonal or having trend. The W-INGARCH-NB(1, 1) model with MEI as covariate has been found outstanding performance in forecasting number of infectious diarrheal patients.
Description:
This Thesis is Submitted to the Institute of Environmental Science (IES), University of Rajshahi, Rajshahi, Bangladesh for The Degree of Doctor of Philosophy (PhD)