Yue Zhao

Thanks for the note. the doc link is fixed :) https://pyhealth.readthedocs.io/en/latest/

if my memory is correct...the data is generated by https://raw.githubusercontent.com/yzhao062/PyHealth/master/examples/data_generation/dataloader_mimic_demo_parallel.py So X includes the features below: var_list = ['Capillary refill rate', 'Diastolic blood pressure', 'Fraction inspired oxygen', 'Glascow coma scale...

Hey Simon. Thanks for reaching out. It is currently at a very early stage. If you are interested in contributing, we may discuss offline for something. Or you are welcome...

thanks for checking. since for now we still need tf and torch for certain models, we would only list these dependencies for reference other than providing a setup or pip...

Dear Prof. Schubert, thanks for the note. Big fan of your work, e.g., ELKI and DAMI evaluation. We fully agree with the quality issue(s) in current outlier research, although among...

That is a great point! That is exactly why we pick 5% as a threshold for new datasets (although they are still adapted). Regarding the absence of more realistic datasets,...

Thank you for your interest. We are working on providing details, e.g., original papers, of each dataset in the next paper release (around late aug to mid-sep). We will ping...

理论上stacking这里的input应该是你已经define好的classifiers，举个例子： ``` clf1 = XGBClassifier(max_depth=3, learning_rate=0.1) clf2 = XGBClassifier(max_depth=5, learning_rate=0.05) clfs = [clf1, clf2] clf_meta = Stacking(clfs) ``` 不知道这个是不是你想问的问题

it should be addressed by latest pyod==0.8.7

that was weired since basically all these methods are calling some operation with axis 1 using np `return np.mean(scores, axis=1).ravel()`