Measurement of dispersion of PM 2.5 in Thailand using confidence intervals for the coefficient of variation of an inverse Gaussian distribution

PeerJ

article

Wasana Chankham¹  Sa-Aat Niwitpong¹  Suparat Niwitpong¹ 

[1] Department of Applied Statistics, King Mongkut’s University of Technology North Bangkok

关键词: Air pollution;    Particulate matter;    Pollution data;    Adjusted generalized confidence interval;    Bootstrap percentile confidence interval;    Fiducial confidence interval;    Fiducial highest posterior;   

DOI  :  10.7717/peerj.12988

学科分类：社会科学、人文和艺术（综合）

来源: Inra

【 摘 要 】

Air pollution is a growing concern for the general public in Thailand with PM 2.5 (particulate matter ≤ 2.5 µm) having the greatest impact on health. The inverse Gaussian (IG) distribution is used for examining the frequency of high concentration events and has often been applied to analyze pollution data, with the coefficient of variation (CV) being used to calculate the quantitative difference in PM 2.5 concentrations. Herein, we propose confidence intervals for the CV of an IG distribution based on the generalized confidence interval (GCI), the adjusted generalized confidence interval (AGCI), the bootstrap percentile confidence interval (BPCI), the fiducial confidence interval (FCI), and the fiducial highest posterior density confidence interval (F-HPDCI). The performance of the proposed confidence intervals was evaluated by using their coverage probabilities and average lengths from various scenarios via Monte Carlo simulations. The simulation results indicate that the coverage probabilities of the AGCI and FCI methods were higher than or close to the nominal level in all of test case scenarios. Moreover, FCI outperformed the others for small sample sizes by achieving the shortest average length. The efficacies of the confidence intervals were demonstrated by using PM 2.5 data from the Din Daeng and Bang Khun Thian districts in Bangkok, Thailand.

【 授权许可】

CC BY