Universiti Teknologi Malaysia Institutional Repository

Modelling asthma cases using count analysis approach: poisson INGARCH and negative binomial INGARCH

Jamaludin, Aaishah Radziah and Yusof, Fadhilah and Suhartono, Suhartono (2020) Modelling asthma cases using count analysis approach: poisson INGARCH and negative binomial INGARCH. MATEMATIKA, 36 (1). pp. 15-30. ISSN 0127-9602

[img]
Preview
PDF
429kB

Official URL: https://dx.doi.org/10.11113/matematika.v36.n1.1158

Abstract

Pollution in Johor Bahru is an issue that needs adequate attention because it has contributed to a number of asthma cases in the area. Therefore, the goal of this study is to investigate the behaviour of asthma disease in Johor Bahru by count analysis approaches namely; Poisson Integer Generalized Autoregressive Conditional Heteroscedasticity (Poisson-INGARCH) and Negative Binomial INGARCH (NB-INGARCH) with identity and log link function. The estimation of the parameter was done by quasi-maximum likelihood estimation. Model assessment was evaluated from the Pearson residuals, cumulative periodogram, the probability integral transform (PIT) histogram, log-likelihood value, Akaike’s Information Criterion (AIC) and Bayesian information criterion (BIC). Our result shows that NB-INGARCH with identity and log link function is adequate in representing the asthma data with uncorrelated Pearson residuals, higher in log likelihood, the PIT exhibits normality yet the lowest AIC and BIC. However, in terms of forecasting accuracy, NB-INGARCH with identity link function performed better with the smaller RMSE (8.54) for the sample data. Therefore, NB- INGARCH with identity link function can be applied as the prediction model for asthma disease in Johor Bahru. Ideally, this outcome can assist the Department of Health in executing counteractive action and early planning to curb asthma diseases in Johor Bahru.

Item Type:Article
Uncontrolled Keywords:Asthma cases, Pollution, Count data, Poisson INGARCH, NB-INGARCH
Subjects:Q Science > QA Mathematics
Divisions:Science
ID Code:85554
Deposited By: Fazli Masari
Deposited On:30 Jun 2020 16:50
Last Modified:30 Jun 2020 16:50

Repository Staff Only: item control page