On the Normal Approximation of the Ratio of Means Estimation of Lognormal Distributions with Application to PM2.5 Concentrations in Northern Thailand.

In this article, we present the point estimation procedures of the ratio of means of two independent lognormal distributions and investigate its accuracy properties. We apply the classical normal approximation procedure and derive the mean and variance parameters of the limiting gaussian distribution. The main criteria of the accuracy are the Bias and Mean Squared Error based on the Monte Carlo simulations. The PM2.5 datasets from two areas are used to illustrate the proposed method, which are consistent with our simulation results. Air pollution measurements of PM2.5 in Northern Thailand from January to April 2022 are examined; they have been identified as the most severe air pollution problem harming the health of people resident in the area. Within a given site, the PM2.5 datasets often follow a right-skewed distribution and are usually fitted by the lognormal model. The mean parameter is used to compute the average mass concentration of PM2.5 at the site and compare it in two different areas. [ABSTRACT FROM AUTHOR]