LOF supporting deletion of arbitrary previous training point

[leejunu]

I see from the documentation that the Local Outlier Factor is meant to support fast insertion/deletion updates, but as far as I can tell, there are no explicit methods for the deletion part. I'm looking to be able to train a LOF over n points, and then delete a user-specified training point (for example, maybe X_10 out of X_1...X_200), then updating the underlying KNN and its metrics. Is this possible with the given methods, in a constant-time manner?