OSDT
OSDT copied to clipboard
xiyanghu
Optimal Sparse Decision Trees
By running some tests I used to get an output like this: [(37,), (-37, 38), (-38, -37, -1), (-38, -37, 1)] How to decode them? I did assume that these...
can you share how to understand tree encoding logic for example what does it mean: leaves_c [(1,), (-7, -1), (-3, -1, 7), (-12, -1, 3, 7), (-1, 3, 7, 12)]...
can you clarify if rule can be extracted as in corels? there is some code https://github.com/xiyanghu/OSDT/blob/master/src/rule.py is it relevant to rule extraction?
can you please share some simple, but meaningful toy example like pycorels to make sure it works for example like https://github.com/fingoldin/pycorels Toy dataset (See picture example above) from corels import...
After installing dependent packages in debian and running test_accuracy.py, I get the following error: Traceback (most recent call last): File "test_accuracy.py", line 107, in file_CART=r'./accuracy/cart_compas.txt', file_OSDT=r'./accuracy/osdt_compas.txt') File "test_accuracy.py", line 67,...
COMPAS contains arithmetic features, how did you binarize those? Were the thresholds manually selected? Was binning used, and if so, how many bins were used and what strategy was used...
