fdatest

The goal of fdatest is to implement various statistical methods for domain selection in functional data analysis, that is selecting a subset of the domain where the difference between two populations is significant. The package is based on the paper by Abramowicz et al. (2022) and Pini & Vantini (2017).

Installation

You can install the package from CRAN with:

install.packages("fdatest")

Alternatively, You can install the development version of fdatest from GitHub with:

# install.packages("pak")
pak::pak("permaverse/fdatest")

Example

The package provides several methods for domain selection, including:

• FDR2: False Discovery Rate for two populations.
• Global2: Global testing for two populations.
• IWT2: Interval-Wise Testing for two populations.
• PCT2: Partition Closed Testing for two populations.
• TWT2: Threshold Wise Testing for two populations.

You can use these methods to test for significant differences between two populations of functional data. Here is an example using the TWT2 method on the NASA temperatures data set:

library(fdatest)

# Performing the TWT for two populations on the NASA temperatures data set
withr::with_seed(1234, {
  out <- TWT2(NASAtemp$paris, NASAtemp$milan)
})
#> Warning: `TWT2()` was deprecated in fdatest 0.2.0.
#> ℹ Use twt2() instead. Be mindful that the argument `statistic` has been
#>   replaced by `aggregation_strategy` and `standardize`.
#> This warning is displayed once per session.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.

# Plotting the results of the TWT
plot(
  out,
  xrange = c(0, 12),
  main = "TWT results for testing mean differences"
)