TreeEnsembleRegressor - 1 vs 3

Next section compares an older to a newer version of the same operator after both definition are converted into markdown text. Green means an addition to the newer version, red means a deletion. Anything else is unchanged.

Files changed (1) hide show

1. TreeEnsembleRegressor1 → TreeEnsembleRegressor3 +0 -12

TreeEnsembleRegressor1 → TreeEnsembleRegressor3 RENAMED Viewed

	@@ -1 +1 @@
1 1	Tree Ensemble regressor. Returns the regressed values for each input in N.
2 2	All args with nodes_ are fields of a tuple of tree nodes, and
3 3	it is assumed they are the same length, and an index i will decode the
4 4	tuple across these inputs. Each node id can appear only once
5 5	for each tree id.
6 6	All fields prefixed with target_ are tuples of votes at the leaves.
7 7	A leaf may have multiple votes, where each vote is weighted by
8 8	the associated target_weights index.
9	- All fields ending with <i>_as_tensor</i> can be used instead of the
10	- same parameter without the suffix if the element type is double and not float.
11 9	All trees must have their node ids start at 0 and increment by 1.
12 10	Mode enum is BRANCH_LEQ, BRANCH_LT, BRANCH_GTE, BRANCH_GT, BRANCH_EQ, BRANCH_NEQ, LEAF
13 11	**Attributes**
14 12	* **aggregate_function**:
15 13	Defines how to aggregate leaf values within a target. <br>One of
16 14	'AVERAGE,' 'SUM,' 'MIN,' 'MAX.'
17 15	* **base_values**:
18 16	Base values for classification, added to final class score; the size
19 17	must be the same as the classes or can be left unassigned (assumed
20 18	0)
21	- * **base_values_as_tensor**:
22	- Base values for classification, added to final class score; the size
23	- must be the same as the classes or can be left unassigned (assumed
24	- 0)
25 19	* **n_targets**:
26 20	The total number of targets.
27 21	* **nodes_falsenodeids**:
28 22	Child node if expression is false
29 23	* **nodes_featureids**:
30 24	Feature id for each node.
31 25	* **nodes_hitrates**:
32	- Popularity of each node, used for performance and may be omitted.
33	- * **nodes_hitrates_as_tensor**:
34 26	Popularity of each node, used for performance and may be omitted.
35 27	* **nodes_missing_value_tracks_true**:
36 28	For each node, define what to do in the presence of a NaN: use the
37 29	'true' (if the attribute value is 1) or 'false' (if the attribute
38 30	value is 0) branch based on the value in this array.<br>This
39 31	attribute may be left undefined and the defalt value is false (0)
40 32	for all nodes.
41 33	* **nodes_modes**:
42 34	The node kind, that is, the comparison to make at the node. There is
43 35	no comparison to make at a leaf node.<br>One of 'BRANCH_LEQ',
44 36	'BRANCH_LT', 'BRANCH_GTE', 'BRANCH_GT', 'BRANCH_EQ', 'BRANCH_NEQ',
45 37	'LEAF'
46 38	* **nodes_nodeids**:
47 39	Node id for each node. Node ids must restart at zero for each tree
48 40	and increase sequentially.
49 41	* **nodes_treeids**:
50 42	Tree id for each node.
51 43	* **nodes_truenodeids**:
52 44	Child node if expression is true
53 45	* **nodes_values**:
54 46	Thresholds to do the splitting on for each node.
55	- * **nodes_values_as_tensor**:
56	- Thresholds to do the splitting on for each node.
57 47	* **post_transform**:
58 48	Indicates the transform to apply to the score. <br>One of 'NONE,'
59 49	'SOFTMAX,' 'LOGISTIC,' 'SOFTMAX_ZERO,' or 'PROBIT'
60 50	* **target_ids**:
61 51	The index of the target that each weight is for
62 52	* **target_nodeids**:
63 53	The node id of each weight
64 54	* **target_treeids**:
65 55	The id of the tree that each node is in.
66 56	* **target_weights**:
67	- The weight for each target
68	- * **target_weights_as_tensor**:
69 57	The weight for each target
70 58	**Inputs**
71 59	* **X** (heterogeneous) - **T**:
72 60	Input of shape [N,F]
73 61	**Outputs**
74 62	* **Y** (heterogeneous) - **tensor(float)**:
75 63	N classes
76 64	**Type Constraints**
77 65	* **T** in (
78 66	tensor(double),
79 67	tensor(float),
80 68	tensor(int32),
81 69	tensor(int64)
82 70	):
83 71	The input type must be a tensor of a numeric type.