Air pollution is a problem faced by most countries across the globe. The modeling and evaluation of the probabilistic behavior of air pollution are crucial in providing useful information that can help in managing the environmental risk and planning for the adverse effects of air pollution. For modeling of air pollution, several statistical approaches have been considered; however, only a few approaches have been used for addressing the uncertainty in the analysis. This study proposes a new application of the Markov chain-based fuzzy states (MCFS) model using triangular fuzzy numbers for analysing the uncertainty in the occurrence of air pollution events and describing the transition behaviour of air pollution. In this study, the air pollution index (API) data collected from the city of Klang in Malaysia for a period between 2012 and 2014 is considered in the analysis. Based on the API data, a five-state Markov chain is considered for representing the five fuzzy states of air pollution. The fuzzy transition probabilities are estimated and used to determine the characteristics of air pollution such as the steady state probabilities and the mean first passage time for each state of air pollution. The findings show that, in general, the risk of occurrences for unhealthy events in Klang is small, nonetheless remains notably troubling. The results demonstrate that the MCFS can effectively model the air pollution index and it could be a better option in predicting air pollution. It may provide valuable information and more understanding about the dynamics of air pollution to the experts and policymakers. This will enable them to develop proper strategies to manage air quality.

空气污染是全球大多数国家面临的问题。空气污染概率行为的建模和评估对于提供有助于管理环境风险和规划空气污染不利影响的有用信息至关重要。对于空气污染建模，已经考虑了几种统计方法；然而，只有少数方法被用于解决分析中的不确定性。本研究提出了基于马尔可夫链的模糊状态 (MCFS) 模型的新应用，该模型使用三角模糊数来分析空气污染事件发生的不确定性并描述空气污染的过渡行为。在这项研究中，分析中考虑了 2012 年至 2014 年期间从马来西亚巴生市收集的空气污染指数 (API) 数据。基于API数据，考虑用五态马尔可夫链来表示空气污染的五种模糊状态。模糊转移概率被估计并用于确定空气污染的特征，例如空气污染的稳态概率和平均首次通过时间。调查结果显示，总的来说，巴生发生不健康事件的风险很小，但仍然令人不安。结果表明，MCFS 可以有效地对空气污染指数进行建模，并且可以成为预测空气污染的更好选择。它可以为专家和政策制定者提供有价值的信息和关于空气污染动态的更多理解。这将使他们能够制定适当的策略来管理空气质量。