Chemometric Treatment of Missing Elements in Air Quality Data Sets
A. Smoliński1, S. Hławiczka2
More details
Hide details
1Central Mining Institute, Department of Energy Saving and Air Protection, Plac Gwarków 1, 40-166 Katowice, Poland 2Institute for Ecology of Industrial Areas, ul. 6 Kossutha, 40-844 Katowice, Poland
Pol. J. Environ. Stud. 2007;16(4):613–622
The article reports the results of an exploratory analysis of an air monitoring data set, collected at a monitoring station in the biggest, most congested and most polluted city of the silesian region, Katowice. In order to extract important information on air pollution in this city, the strategy of exploring the data set with missing elements and outliers simultaneously existing in the data was used. The strategy assumed the initial estimation of missing elements based on the application of robust Partial Least Squares (rPLS) and outliers identification based on the so-called robust distance. After outliers identification and replacing them with missing elements, the Expectation-Maximization iterative approach (built into Principal Component Analysis (PCA)) was used for the construction of the final model.