In response to #28, this PR implements the matrix-based Frobenius loss function. I need to benchmark cvCovEst's performance under each loss before merging PR.
@nhejazi What are your thoughts on this? I think it has potential to confuse some future readers/users of the method, but it might be worth doing nonetheless if this matrix-based loss is more efficient. Perhaps we can get away with it if we emphasize their equivalence in terms of selection in the manuscript?
In response to #28, this PR implements the matrix-based Frobenius loss function. I need to benchmark
cvCovEst
's performance under each loss before merging PR.@nhejazi What are your thoughts on this? I think it has potential to confuse some future readers/users of the method, but it might be worth doing nonetheless if this matrix-based loss is more efficient. Perhaps we can get away with it if we emphasize their equivalence in terms of selection in the manuscript?