How trees split and build partitions.
0/10
completed
A decision tree consistently selects high-cardinality categorical features (e.g., zip code with 1000 values) near the root of the tree. What bias does this introduce?