run_fpca function

run_fpca

This is a wrapper for the function fpca.face from the refund package. EXPAND

run_fpca(
  mxFDAobject,
  metric = "uni k",
  r = "r",
  value = "fundiff",
  knots = NULL,
  analysis_vars = NULL,
  lightweight = FALSE,
  filter_cols = NULL,
  ...
)

Arguments

• mxFDAobject: object of class mxFDA created by make_mxfda with metrics derived with extract_summary_functions
• metric: name of calculated spatial metric to use
• r: Character string, the name of the variable that identifies the function domain (usually a radius for spatial summary functions). Default is "r".
• value: Character string, the name of the variable that identifies the spatial summary function values. Default is "fundiff".
• knots: Number of knots for defining spline basis.Defaults to the number of measurements per function divided by 2.
• analysis_vars: Optional list of variables to be retained for downstream analysis.
• lightweight: Default is FALSE. If TRUE, removes Y and Yhat from returned FPCA object. A good option to select for large datasets.
• filter_cols: a named vector of factors to filter summary functions to in c(Derived_Column = "Level_to_Filter") format
• ...: Optional other arguments to be passed to fpca.face

Returns

A mxFDA object with the functional_pca slot filled for the respective spatial summary function containing: - mxfundata: The original dataframe of spatial summary functions, with scores from FPCA appended for downstream modeling

• fpc_object: A list of class "fpca" with elements described in the documentation for refund::fpca.face

Details

The filter_cols parameter is useful when the summary function was input by the user using add_summary_function() and the multiple marks were assessed; a column called "Markers" with tumor infiltrating lymphocytes as well as cytotoxic T cells. This parameter allows for filtering down to include only one or the other.

Examples

#load ovarian mxFDA object
data('ovarian_FDA')

#run the FPCA
ovarian_FDA = run_fpca(ovarian_FDA, metric = "uni g", r = "r", value = "fundiff",
                       lightweight = TRUE,
                       pve = .99)

References

Xiao, L., Ruppert, D., Zipunnikov, V., and Crainiceanu, C. (2016). Fast covariance estimation for high-dimensional functional data. Statistics and Computing, 26, 409-421. DOI: 10.1007/s11222-014-9485-x.

Author(s)

Julia Wrobel julia.wrobel@emory.edu

Alex Soupir alex.soupir@moffitt.org