A synthetic air pollution index using a T2 approach
Keywords:three-way analysis, principal component analysis, T2-Tucker model, air quality index
It is generally acknowledged that exceedance of the safe thresholds of air pollutants significantly affect human health. Hence, we can understand the importance of having a synthetic measure for overall air pollution, in a certain geographical area, computable using an index: to the values taken on by this index are then associated different health risks. In order to compute a reliable and comparable air quality index, in this paper we suggest an analytical procedure based on data reduction functions for summarizing three-way arrays, here consisting of the observations on a few major pollutants monitored over time at multiple sites. These functions reduce the dimensions of the array by applying a joint principal component analysis. Symmetrical three-way PCA has been successfully applied in environmental experimental studies, where the three dimensions of the data array are typically of similar length. Instead, a long time dimension usually characterizes environmental pollution data, where an asymmetrical twoway PCA is preferable: the aim of this paper is to indicate an approach that, by reducing two of the three dimensions, leads to the computation of an informative air quality index. Plaia et al. (2013) have applied an analogous approach, but reducing just one dimension of the array.
Copyright (c) 2020 Giuliana Passamani, Paola Masotti
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.