This is an R package that provide a way to analyze spatial-temporal data based on SSTL (Spatial Seasonal Trend decomposition using Loess) routine under the divide and recombined framework.
# from github
devtools::install_github("XiaosuTong/drsstl")
The main function to call in the package is drsstl
function, which is a
wrapper function of sstl_mr
and sstl_local
two functions. drsstl
can
be used to carry out SSTL routine on the dataset either in the memory or
on the HDFS.
The function for prediction is predNewLocs
function, which is a wrapper
function of predNew_mr
and predNew_local
two functions. Depends on where
the original fitting results are saved, the prediction at new locations
will be conducted either in local memory or on HDFS using MapReduce.
In the data subdirectory, tmax.RData
inlcudes a data.frame named tmax_all
which can be used as example to illustrate the usage of drsstl
in local memory.
station_info.RData
contains a data.frame which includes all stations information.
It is usefual for both local and mapreduce situation of drsstl
. User has to
copy the station_info.RData
to HDFS if the mapreduce is chosen.
In inst subdirectory, tmax.txt is the raw data file contains the maximum
temperature data. User can copy it to HDFS and conduct drsstl
under divide
and recombined framework.
drsstl development is depended on
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.