makeBmrsExtraction: Compute and extract relevant information from a batch of...

In pokyah/agrometeoR: Spatial interpolation for Agromet project @ CRA-W

Description

Compute and extract relevant information from a batch of benchmark experiments. This extraction can then be used to produce various plots.

Usage

makeBmrsExtraction(benchmarkResult, tasks, tasks_sids, as.df = TRUE)

Arguments

benchmarkResult	an object of class mlr::benchmark() produced by agrometeoR::makeBrmsBatch
tasks	a list which elements are of class mlr::makeRegrTask(). These tasks must be the same that were passed to makeBmrsBatch
tasks_sids	a list which elements are integer vectors containing the sids of the used stations of each task. This list is provided as an output element of the agrometeoR::makeTaks() function.
as.df	a boolean specifying if the result must be returned as a single dataframe rather than list of lists. Default is TRUE

Value

A 2 elements named list

• snitch : a boolean. Is TRUE if function has provided the expected result. Is FALSE is function throws an error

• output : a named list which elements are :

  • value : an element of class data.frame if parameter as.df is set to TRUE. If set to FALSE, the function returns a list which elements are dataframes

  • condition : a character specifying the condition encountered by the function : success, warning, or error.

  • message : a character specifying the message relative to the condition.

Author(s)

Thomas Goossens

Examples

## Not run: 
# load magrittr for pipe use : %>%
library(magrittr)

# create the dataset
myDataset = makeDataset(
  dfrom = "2017-03-04T15:00:00Z",
  dto = "2017-03-04T18:00:00Z",
  sensor = "tsa")

# extract the list of hourly sets of records
myDataset = myDataset$output$value

# create the tasks
myTasks = purrr::map(myDataset, makeTask, target = "tsa")

# extract the used sids of each task from the outputs
myUsedSids = myTasks %>% purrr::modify_depth(1, ~.$output$stations$used)

# extract the tasks from the outputs
myTasks = myTasks %>% purrr::modify_depth(1, ~.$output$value$task)

# Conduct a batch of benchmarks experiments without saving temp files
myBmrsBatch = makeBmrsBatch(
tasks = myTasks,
learners = agrometeorLearners,
measures = list(mlr::rmse),
keep.pred = TRUE,
models = FALSE,
groupSize  = NULL,
level = "mlr.benchmark",
resamplings = "LOO",
cpus = 1,
prefix = NULL,
temp_dir = NULL,
removeTemp = FALSE)

# Keep the relevant information
myBmrsBatch = myBmrsBatch$output$value

# Get the extraction from myBmrsBatch
myBmrsExtraction = makeBmrsExtraction(myBmrsBatch, myTasks, myUsedSids, as.df = TRUE)

# Keeping the relevant information
myBmrsExtraction = myBmrsExtraction$output$value

# Get an excerpt of the output
head(myBmrsExtraction)

## End(Not run)

pokyah/agrometeoR documentation built on May 26, 2019, 7 p.m.