search

RektPunk / MQBoost

Multiple quantiles estimation model maintaining non-crossing condition (or monotone quantile condition) using LightGBM and XGBoost
24 stars 4 forks source link

Example doesn't work / Plans for categorical data? #25

Open fleicher opened 2 weeks ago

fleicher commented 2 weeks ago

Hey, I am trying to run the examples.

But when I am executing mqoptimizer.py, I get the error that the parameter epsilon when creating the object MQObjective is not set. A default value should solve the problem.

mq_optimizer.optimize_params(
    dataset=train_dataset,
    n_trials=10,
)

Also, one question, do you have any plans to natively support categorical data (at least for LightGBM)?

Thank you for your work

RektPunk commented 2 weeks ago

Hey, @fleicher. Thank you so, so much for letting me know. You are literally my lifesaver! My response is a bit late, but I fixed it quickly as soon as I checked! :) BTW, could you explain a bit more what you mean by 'natively support categorical data' at the end?

RektPunk commented 1 week ago

Hey @fleicher, I'm working on 'support categorical data' at https://github.com/RektPunk/MQBoost/pull/32. Hope it is what you expected.