TreeEnsembleRegressor - 1 vs 3

Files changed (1) hide show
  1. TreeEnsembleRegressor1 → TreeEnsembleRegressor3 +12 -0
TreeEnsembleRegressor1 → TreeEnsembleRegressor3 RENAMED
@@ -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.
9
11
  All trees must have their node ids start at 0 and increment by 1.
10
12
  Mode enum is BRANCH_LEQ, BRANCH_LT, BRANCH_GTE, BRANCH_GT, BRANCH_EQ, BRANCH_NEQ, LEAF
11
13
  **Attributes**
12
14
  * **aggregate_function**:
13
15
  Defines how to aggregate leaf values within a target. <br>One of
14
16
  'AVERAGE,' 'SUM,' 'MIN,' 'MAX.'
15
17
  * **base_values**:
16
18
  Base values for classification, added to final class score; the size
17
19
  must be the same as the classes or can be left unassigned (assumed
18
20
  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)
19
25
  * **n_targets**:
20
26
  The total number of targets.
21
27
  * **nodes_falsenodeids**:
22
28
  Child node if expression is false
23
29
  * **nodes_featureids**:
24
30
  Feature id for each node.
25
31
  * **nodes_hitrates**:
32
+ Popularity of each node, used for performance and may be omitted.
33
+ * **nodes_hitrates_as_tensor**:
26
34
  Popularity of each node, used for performance and may be omitted.
27
35
  * **nodes_missing_value_tracks_true**:
28
36
  For each node, define what to do in the presence of a NaN: use the
29
37
  'true' (if the attribute value is 1) or 'false' (if the attribute
30
38
  value is 0) branch based on the value in this array.<br>This
31
39
  attribute may be left undefined and the defalt value is false (0)
32
40
  for all nodes.
33
41
  * **nodes_modes**:
34
42
  The node kind, that is, the comparison to make at the node. There is
35
43
  no comparison to make at a leaf node.<br>One of 'BRANCH_LEQ',
36
44
  'BRANCH_LT', 'BRANCH_GTE', 'BRANCH_GT', 'BRANCH_EQ', 'BRANCH_NEQ',
37
45
  'LEAF'
38
46
  * **nodes_nodeids**:
39
47
  Node id for each node. Node ids must restart at zero for each tree
40
48
  and increase sequentially.
41
49
  * **nodes_treeids**:
42
50
  Tree id for each node.
43
51
  * **nodes_truenodeids**:
44
52
  Child node if expression is true
45
53
  * **nodes_values**:
46
54
  Thresholds to do the splitting on for each node.
55
+ * **nodes_values_as_tensor**:
56
+ Thresholds to do the splitting on for each node.
47
57
  * **post_transform**:
48
58
  Indicates the transform to apply to the score. <br>One of 'NONE,'
49
59
  'SOFTMAX,' 'LOGISTIC,' 'SOFTMAX_ZERO,' or 'PROBIT'
50
60
  * **target_ids**:
51
61
  The index of the target that each weight is for
52
62
  * **target_nodeids**:
53
63
  The node id of each weight
54
64
  * **target_treeids**:
55
65
  The id of the tree that each node is in.
56
66
  * **target_weights**:
67
+ The weight for each target
68
+ * **target_weights_as_tensor**:
57
69
  The weight for each target
58
70
  **Inputs**
59
71
  * **X** (heterogeneous) - **T**:
60
72
  Input of shape [N,F]
61
73
  **Outputs**
62
74
  * **Y** (heterogeneous) - **tensor(float)**:
63
75
  N classes
64
76
  **Type Constraints**
65
77
  * **T** in (
66
78
  tensor(double),
67
79
  tensor(float),
68
80
  tensor(int32),
69
81
  tensor(int64)
70
82
  ):
71
83
  The input type must be a tensor of a numeric type.