Principal Components Analysis of a matrix using Non-linear Iterative Partial Least Squares or weighted Expectation Maximization PCA with Gram-Schmidt orthogonalization of the scores and loadings. Optimized for speed. See Andrecut (2009)
Minor fix: add row/column names to fitted matrix.
New: startcol argument can now be a function.
Slight change to automatic start column selection. By default, the start column is the column with the largest sum of absolute values. (Formerly was largest variance.)
The nipals function was split off from the gge package, extensively optimized and compared with implementations in other packages.
Find published example of NIPALS with missing data. (Only found 1)