parTemporalBImp

<p>Prediction dataset including many mssing values</p>

allPre_

siteids_

isite_

pol_season_trends_

<p>A function to use constraint pptimization to predict the missing values.</p>

 imputation 

The approach of constrained spatiotemporal mixed models is to make reliable estimation of air pollutant concentrations at high spatiotemporal
resolution (Li, L., Zhang, J., Meng, X., Fang, Y., Ge, Y., Wang, J., Wang, C., Wu, J., Kan, H. (2018) <doi.org/10.1016/j.rse.2018.09.001>;
Li, L., Lurmann, F., Habre, R., Urman, R., Rappaport, E., Ritz, B., Chen, J., Gilliland, F., Wu, J., (2017) <doi:10.1021/acs.est.7b01864>).
This package is an extensive tool for this modeling approach with support of block Kriging (Goovaerts, P. (1997) <http://www.gbv.de/dms/goettingen/229148123.pdf>)
and uses the PM2.5 modeling as examples. It provides the following functionality:
(1) Extraction of covariates from the satellite images such as GeoTiff and NC4 raster;
(2) Generation of temporal basis functions to simulate the seasonal trends in the study regions;
(3) Generation of the regional monthly or yearly means of air pollutant concentration;
(4) Generation of Thiessen polygons and spatial effect modeling;
(5) Ensemble modeling for spatiotemporal mixed models, supporting multi-core parallel computing;
(6) Integrated predictions with or without weights of the model's performance, supporting multi-core parallel computing;
(7) Constrained optimization to interpolate the missing values;
(8) Generation of the grid surfaces of air pollutant concentration estimates at high resolution;
(9) Block Kriging for regional mean estimation at multiple scales.

parTemporalBImp: Function to Fill Missing Values by Constraint Optimization

Description

Usage

Arguments

Value