| Summary: | 碩士 === 逢甲大學 === 應用數學系 === 105 === Air pollution is one of the most important global environmental issues. Taiwan Environmental Protection Agency is currently using Air Quality Index (AQI) to measure air quality. It indicates how air quality influences the health of sensitive groups and all other people when AQI exceeds 100 and 150, respectively. The objective of this study is to assess the air quality per hour every month in Taichung from 2014 to 2016 based on the lower confidence limits for nonconformance proportion of AQI. The nonconformance proportion is defined as the proportion when a characteristic of interest falls out an acceptance region of a specification. This study collects concentration data of individual air pollutants from five air quality monitoring sites in Taichung from 2014 to 2016 and calculates the real-time AQI. Next, an unbalanced two-way random effects model for AQI is constructed and the nonconformance proportions of AQI are defined. Finally, we use three kinds of estimation methods including the t based method, the adjusted t-based method and the generalized pivotal quantity based method. They are conducted to simulate and calculate the lower confidence limits for nonconformance proportion of AQI and evaluate their performance. The results indicate that the air quality in Taichung has greater impact on human health in March and October. In general, using the adjusted t-based method to estimate the lower confidence limits of nonconformance proportion for AQI is better and stable. However, the three estimation methods are unstable when nonconformance proportion for AQI is lower than 0.2. This is the topic of further researches in the future.
|
|---|
