Function impute_missing_values 

pub fn impute_missing_values(
    data: &FdMatrix,
    argvals: &[f64],
    method: ImputationMethod,
) -> Result<FdMatrix, FdarError>

Impute NaN values in a regular functional data matrix.

Returns a new FdMatrix with NaN entries replaced according to method. Non-NaN entries are copied through unchanged.

For Linear, gaps at the curve boundary (no neighbor on one side) are filled with the nearest valid value (boundary extension).

Arguments

• data — Functional data matrix (n × m) with possible NaN entries
• argvals — Evaluation points (length m, must be sorted)
• method — Imputation strategy

Errors

• FdarError::InvalidDimension if argvals.len() != data.ncols()
• FdarError::InvalidParameter if any curve consists entirely of NaN values