Modeling and Forecasting Indian Malaria Incidence Using Generalized Time Series Models. Issue 2 (3rd April 2019)

• Record Type: Journal Article
• Authors: Mukhopadhyay, S.; Tiwari, R.; Shetty, P.; Gogtay, N.J.; Thatte, U.M.
• Is Part Of: Communication in statistics. Volume 5:Issue 2(2019)
• Page Start: 111
• Page End: 120
• Publication Date: 2019-04-03
• Subjects: ARMA -- generalized linear models -- forecasting -- negative binomial -- overdispersion; Mathematical statistics -- Data processing -- Periodicals; 519.505
• DOI: 10.1080/23737484.2019.1580629
• Languages: English
• ISSNs: 2373-7484

Abstract

Abstract: In this article, time series for generalized linear models (GLMs) is used to fit weekly malaria data collected from the largest tertiary care hospital in Mumbai, India. Since the data are positive counts of malaria, conditional distribution of the disease counts given the past outcomes and the covariates are assumed to be distributed as a Poisson or a negative binomial. The data analysed show a downward trend, with the presence of seasonality and overdispersion. Various time series models for GLMs are compared and the best model is selected for forecasting the malaria counts. Forecasting regions and model fitting statistics are reported.