Saturday, April 26, 2014

PM2.5 Map by fusing Machine-learning and Kriging estimates

Just a brief update on our progress in making PM2.5 maps for the northeast. First we applied machine learning algorithms to estimate PM2.5 from remote sensing, ground station and meteorology data, then we fused Kriging results of the ground station data to obtain the final PM2.5 map. Inverse distance weighting on remote sensing has been applied to improve the coverage on remote sensing. The results were obtained using NY state data as we were funded by NY state agency.