Applications of Poisson Hidden Markov models to PM10 concentrations data

• 영문명: Applications of Poisson Hidden Markov models to PM10 concentrations data
• 발행기관: 한국자료분석학회
• 저자명: Hee-Young Kim
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.24 No.4, 1203~1212쪽, 전체 10쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2022.08.31

영문 초록

This study addresses the problem of monitoring and forecasting of particulate matter(PM) data, focusing, in particular, on high-level , which is known to adversely impact human mortality and morbidity. We use hourly data, collected between November 26, 2018, to December 01, 2018, from 40 stations located in the Seoul metropolitan area of South Korea. We model the number of regions corresponding to “bad” or “very bad” categories of the density, using a hidden Markov model with Poisson state-dependent distribution, Poisson-HMM, since a Poisson-HMM allows for both overdispersion and serial dependence. Model selection, in particular for the number of latent states, is based on Akaike’s Information criterion and Bayesian Information criterion. By AIC and BIC, the Poisson-HMMs are a big improvement on independent mixture of Poissons with the same state, which does not allow for the serial dependence in the observations. We conclude that Poisson-HMM is a good forecasting model.

키워드

PM10 Count data Time series modeling Hidden Markov model