The functional data analysis of hourly air pollution data: Canonical correlation and principal component analyses of PM10, PM2.5 and ozone data for El Paso, Texas
This study investigates diurnal patterns and the associations between the daily hourly measurements of air pollutants PM2.5 and Ozone collected over a six year period at the air monitoring stations Cams 12 at 250 Rim Road, and Cams 40 at Sun Metro Station, El Paso. Those days having 5 or more hours of data were considered missing days and omitted from the data analysis. For those days with fewer than 5 hours of missing observations, P-splines were used to reconstruct the curves on the entire interval [0,23]. Then using the functional data software for S Plus pioneered by Prof. Jim Ramsay, principal component analysis and canonical correlation analysis were performed on the data. For the functional data analysis, the preferred regression spline basis functions are the cubic B-splines. A discussion on nonparametric regression and splines is included. On the basis of the results of the analysis, it was found that no significant relationship exists between PM2.5 and Ozone.^
Samuels, Vernon, "The functional data analysis of hourly air pollution data: Canonical correlation and principal component analyses of PM10, PM2.5 and ozone data for El Paso, Texas" (2006). ETD Collection for University of Texas, El Paso. AAI1439480.