tau-mlwell / set-tree Goto Github PK

I'm a graduate student at HKUST. I have read your marvelous article " Trees with Attention for Set Prediction Tasks” in ICML 2021 with great interest. Could you please consider sharing your datasets(including Drug Prescription Errors & Redshift Estimation) with me for academic purposes? I'm in particular interested in these two datasets.

Thanks for your consideration!

Hi,

Thank you for this wonderful tool.

I've been unable to change the minimum samples for leaf nodes. Everything else works fine when generating the trees (I was able to change the maximum depth for example).

This is how I instantiated the tree:
set_tree_model = settree.SetTree(
classifier=False,
criterion="mse",
splitter='sklearn',
max_features=None,
min_samples_split=3,
operations=list_of_operations,
use_attention_set=USE_ATTN_SET,
use_attention_set_comp=USE_ATTN_SET_COMP,
attention_set_limit=ATTN_SET_LIMIT,
max_depth=2,
min_samples_leaf=3,
random_state=SEED)

Thanks again!

Hi,

When trying to get the feature importances of a trained gbest, the following errors occur:

1. line 622 in gbest.py: 'SetSplitNode' object has no attribute 'node_count'
   I tried to fix it by replace the 'node_count' with 'tree.n_nodes'
   but even then, another error appears:

2. When trying to execute:
   relevant_trees[0].tree_.compute_feature_importances(normalize=False)
   we get
   AttributeError: 'SetSplitNode' object has no attribute 'compute_feature_importances'

It looks like SetSplitNode (or rather settree.set_tree.SetTree) does not implement compute_feature_importances method.

Can you suggest any alternative to calculate the feature importances?

Thanks in advance