stlfit: Apply the stlplus fitting at each location in parallel

In XiaosuTong/drsstl: Spatial Seasonal Trend Loess analysis routine using divide and recombined

Description

Calling stlplus function from Ryan Hafen's stlplus package on time series at each location in parallel. Every station uses the same smoothing parameter

Usage

stlfit(input, output, model_control = spacetime.control(),
  cluster_control = mapreduce.control())

Arguments

input	The path of input sequence file on HDFS. It should be by location division.
output	The path of output sequence file on HDFS. It is by location division but with seasonal and trend components
model_control	The list contains all smoothing parameters
cluster_control	A list contains all mapreduce tuning parameters.

Author(s)

Xiaosu Tong

References

R. B. Cleveland, W. S. Cleveland, J. E. McRae, and I. Terpenning (1990) STL: A Seasonal-Trend Decomposition Procedure Based on Loess. Journal of Official Statistics, 6, 3–73.

Ryan Hafen (2010): stlplus: Local regression models: Advancements, applications, and new methods. Purdue University

See Also

spacetime.control, mapreduce.control

Examples

## Not run: 
    FileInput <- "/tmp/bystat"
    FileOutput <- "/tmp/bystatfit"
    ccontrol <- mapreduce.control(libLoc=NULL, reduceTask=0)
    mcontrol <- spacetime.control(
      vari = "resp", time = "date", n = 576, stat_n=7738, n.p = 12, s.window = "periodic",
      t.window = 241, degree = 2, span = 0.015, Edeg = 2
    )
    stlfit(FileInput, FileOutput, model_control=mcontrol, cluster_control=ccontrol)

## End(Not run)

XiaosuTong/drsstl documentation built on May 9, 2019, 11:06 p.m.