pyparpredict: Python parallel computation for prediction

In lspatial/sptemUS: Constrained Spatiotemporal Mixed-Effect Models for Exposure Estimation of Air Pollutants

Description

Python parallel computation for prediction by access to server.

Usage

pyparpredict(dset, flag = "prefile", nModel = 6, ncore = 6)

Arguments

dset	Data frame to be predicted
flag	Flag path to save the predicted files
nModel	Number of models
ncore	Number of cores

Details

This function is to leverage the server to get the compution

Value

Use the flag to get the final predictions.

Author(s)

Lianfa Li

Examples

# This example illustrates how to use the python to call R function to make parallel predictions
#  for NO2 and NOx concentration.  The output of multiple models will be saved on the server side using the flag path.
#   You can use the function, weiA2EnsFlag to get the weighted averages from the saved files using the project flag.

# The first step is to ensure no missing vaues in the prediction dataset

data(sc_sample_m)
colnames=c('dist_fcc1','min_temp', 'winsp', 'popd','cl_freewaynox','cl_nonfreewaynox', 'rid500m',
           'aqs_add_no2_y', 'aqs_add_nox_y','stbasis_no2_1','stbasis_no2_2','stbasis_nox_1','stbasis_nox_2',
           'trdenscaled300_5km_r')

exp="sc_sample_m[which("
for(i in c(1:length(colnames))){
  acol=colnames[i]
  exp=paste(exp,"!is.na(sc_sample_m[,'",acol,"'])",sep="")
  if(i<length(colnames)){
    exp=paste(exp," & ",sep="")
  }
}
exp=paste(exp,"),]",sep="")
print(exp)
sc_sample_sub=eval(parse(text=exp))

# Then, call pyparpredict to initiate the parallel predictions on the server side.

no2_ens=pyparpredict(sc_sample_sub,'no2',flag='sc_sample_no2',nModel=10,ncore=5,idF="gid")
nox_ens=pyparpredict(sc_sample_sub,'nox',flag='sc_sample_nox',nModel=10,ncore=5,idF="gid")

lspatial/sptemUS documentation built on May 29, 2019, 3:42 a.m.