TY - JOUR
TI - Evaluation of land use regression models for NO2 and particulate matter in 20 European study areas: The ESCAPE project
AU - Wang, M.
AU - Beelen, R.
AU - Basagana, X.
AU - Becker, T.
AU - Cesaroni, G.
AU - De Hoogh, K.
AU - Dedele, A.
AU - Declercq, C.
AU - Dimakopoulou, K.
AU - Eeftens, M.
AU - Forastiere, F.
AU - Galassi, C.
AU - Gražulevičiene, R.
AU - Hoffmann, B.
AU - Heinrich, J.
AU - Iakovides, M.
AU - Künzli, N.
AU - Korek, M.
AU - Lindley, S.
AU - Mölter, A.
AU - Mosler, G.
AU - Madsen, C.
AU - Nieuwenhuijsen, M.
AU - Phuleria, H.
AU - Pedeli, X.
AU - Raaschou-Nielsen, O.
AU - Ranzi, A.
AU - Stephanou, E.
AU - Sugiri, D.
AU - Stempfelet, M.
AU - Tsai, M.-Y.
AU - Lanki, T.
AU - Udvardy, O.
AU - Varró, M.J.
AU - Wolf, K.
AU - Weinmayr, G.
AU - Yli-Tuomi, T.
AU - Hoek, G.
AU - Brunekreef, B.
JO - Environmental science and technology
PY - 2013
VL - 47
TODO - 9
SP - 4357-4364
PB - 
SN - 0194-0287
TODO - 10.1021/es305129t
TODO - Absorbances;  Land-use regression models;  Nitrogen dioxides;  Particulate Matter;  Pm concentrations;  Predictive abilities;  Small training;  Study areas, Land use;  Regression analysis, Nitrogen oxides, nitric dioxide;  oxide;  unclassified drug, absorption;  atmospheric chemistry;  concentration (composition);  copper;  data set;  fractionation;  nitrogen dioxide;  numerical model;  particulate matter, absorption;  accuracy;  article;  concentration (parameters);  correlation coefficient;  Europe;  land use;  particle size;  particulate matter;  predictive value;  regression analysis;  statistical model, Air Pollution;  Europe;  Models, Theoretical;  Nitric Oxide;  Particulate Matter, Europe
TODO - Land use regression models (LUR) frequently use leave-one-out-cross- validation (LOOCV) to assess model fit, but recent studies suggested that this may overestimate predictive ability in independent data sets. Our aim was to evaluate LUR models for nitrogen dioxide (NO2) and particulate matter (PM) components exploiting the high correlation between concentrations of PM metrics and NO2. LUR models have been developed for NO2, PM2.5 absorbance, and copper (Cu) in PM10 based on 20 sites in each of the 20 study areas of the ESCAPE project. Models were evaluated with LOOCV and "hold-out evaluation (HEV)" using the correlation of predicted NO2 or PM concentrations with measured NO2 concentrations at the 20 additional NO2 sites in each area. For NO2, PM2.5 absorbance and PM10 Cu, the median LOOCV R2s were 0.83, 0.81, and 0.76 whereas the median HEV R 2 were 0.52, 0.44, and 0.40. There was a positive association between the LOOCV R2 and HEV R2 for PM2.5 absorbance and PM10 Cu. Our results confirm that the predictive ability of LUR models based on relatively small training sets is overestimated by the LOOCV R2s. Nevertheless, in most areas LUR models still explained a substantial fraction of the variation of concentrations measured at independent sites. © 2013 American Chemical Society.
ER -