This can be used to implement custom FDA feature extraction. Takes a learn and a reextract function along with some optional parameters to those as argument.

makeExtractFDAFeatMethod(learn, reextract, args = list(), par.set = NULL)



(function(data, target, col, ...))
Function to learn and extract information on functional column col. Arguments are:

  • data data.frame
    Data.frame containing matricies with one row per observation of a single functional or time series and one column per mea time point. All entries need to be numeric.

  • target (character(1))
    Name of the target variable. Default: “NULL”. The variable is only set to be consistent with the API.

  • col (character(1) | numeric(1))
    column names or indices, the extraction should be performed on. The function has to return a named list of values.


(function(data, target, col, ...))
Function used for reextracting data in predict phase. Can be equal to learn.


Named list of arguments to pass to learn via ....


Paramset added to the learner if used in conjunction with a makeExtractFDAFeatsWrapper. Can be NULL.`

See also