While this code works well and outputs results properly, it takes an incredibly large time to parse large datasets. The dataset in particular that I am using is an NLP dataset, particular of TF values, so there are a LOT of zeroes (not sure if that makes a difference). My dataset's size is (550683, 10891). That is estimated to take more than 10 days to finish on my current hardware.

While this code works well and outputs results properly, it takes an incredibly long time to process large datasets. The dataset in particular that I am using is an NLP dataset, particularly of term frequency values, so there are a LOT of zeroes and the data does not follow a normal distribution (not a single feature does) (not sure if that makes a difference). My dataset's size is (550683, 10891). That is estimated to take more than 10 days to finish on my current hardware.

How can I optimize this code to improve performance? Using the make_multilabel_classification call above, even that takes a fair amount of time given the feature space.

How can I optimize this code to improve performance? Using the make_multilabel_classification call above, even that takes a fair amount of time given the feature space. .

Please note, if you believe the algorithm was implemented incorrectly please feel free to fix that. Principally concerned with speed