fit function completes much faster than predict, is this normal?

[kotoroshinoto]

Was attempting to train a classification model on scikit-eLCS. In addition to not placing any rows in one of the two output classes, I noticed the training finished much faster than the prediction function.

That doesn't smell right to me, but I figured I should ask.

[ryanurbs]

I have noticed some odd behavior with this implementation on certain datasets, but we have not had a chance to track down if there is a bug or this is just part of the limited performance of eLCS as a very basic LCS algorithm. I hope to revisit it in the future, but not entirely sure when that will be.