I like the Nadaraya-Watson (NW) kernel regressor because it allows discrete windowed linear smoothing on non-uniformly sampled data without explicitly presuming a fit function.
It is a "non-parametric" approach and that means it has a different set of assumptions about the basis than typically "parametric" approaches. It " in finite samples the NW estimator tends to have a smaller variance" than expectation-maximization methods.
Is there an estimator built into LabVIEW that has equivalent or greater performance?
Some references: