Estimating Air Pollution Index in Indonesia as Effort to Increase Life Expectancy Fanny Novika
Trisakti School of Insurance
Abstract
High levels of unhealthy Air Pollution Index (API) can severely impact human health and the environment. Thus, effective preparation for such risk events relies on precise estimation of unhealthy API levels. This study proposes the Hierarchical Generalized Pareto Distribution (H-GPD) Method based on the Bayesian framework to obtain accurate estimation of API exceedance in Indonesia. This study will look at the provinces with the worst API as an effort to obtain great benefits if they succeed in cleaning the air effectively so that they can increase life expectancy. To produce the model, three parameters are determined in H-GDP, namely location, scale and shape in each province in Indonesia. Unhealthy API are governed by the Fu distribution function of variable x above the threshold u = 100. The data used in this study are data from IQAir in Indonesia and BMKG. Furthermore, the GDP parameters will be identified and the spatial and seasonal impacts on the marginal density of API data exceedances will be determined using the Hierarchical Model. The accuracy of the model using Goodness of fit through Deviance Information Criteria (DIC) and Akaike^s Information Criteria (AIC). The results show that provinces that have unhealthy API are South Sumatra, West Java, DKI Jakarta, East Java and Central Kalimantan.
Keywords: Air Polution Index- Deviance Information Criteria- Hierarchical Generalized Pareto Distribution- Life Expentancy-
