col_gen_estimator.Row

class col_gen_estimator.Row

To make the processing faster, we store some row related information in this class.

Attributes:

id: (int) ID of the row.

reachable_nodes: (set(int)) IDs of the nodes the row can reach.

reachable_paths = (set(int)) IDs of the paths the row can follow.

reachable_leaves = (set(int)) IDs of the leaves the row can reach.

target: (int) Target of the row.

left_splits: (set(int)) IDs of the splits where the feature value of the

row is smaller or equal to the threshold.

right_splits: (set(int)) IDs of the splits where the feature value of

the row is greater than the threshold.

removed_from_master: (bool) True if the row is removed from the master.

This can happen if the row can only reach at max two leaves or the row is similar to some other row in the dataset with respect to the splits being considered.

removed_from_sp: (bool) True if the row is removed from the master and the

subproblem. This can happen if the row is similar to some other row in the dataset with respect to the splits being considered.

weight: (int) Typically = 1. Some rows can have higher weights if we remove

the other rows that are similar to this row from the dataset.

__init__() → None

Examples using col_gen_estimator.Row

sphx_glr_auto_examples_dtree_experiment.py

Code to launch decision tree experiments.