I’m using ExpressAnalyst for a proteomics analysis and selected kNN (feature-wise, k=10) for missing value imputation. I saw in the documentation that it uses k=10 by default — thanks for making that clear.
For my methods write-up, I just want to confirm: is the kNN imputation single-pass (each missing value estimated once from its k nearest observed neighbors), or does it iterate (re-estimating missing values using previously imputed values in subsequent passes)?
I’m assuming single-pass consistent with the standard impute.knn implementation — just want to verify before I cite it