DecisionTree Syntax Elements - Teradata Vantage

OutputTable

Specify the name for the output table that is to contain the final decision tree (the model table). The output_table must not exceed 64 characters.

FinalResponseTable

[Optional] Specify the name for the output table that is to contain the final PID and response pair from the response table and the node_id from the final single drive tree.

Default behavior: The function does not output this table.

IntermediateSplitsTable

[Disallowed with SplitsTable, optional otherwise.] Specify the name for the intermediate splits table, if it is to be saved.

Default behavior: The function does not save the intermediate splits table.

AttributeNameColumns

Specify the names of the InputTable or AttributeTable columns that have the attribute names.

AttributeValueColumn

Specify the name of the InputTable or AttributeTable columns that have the attribute values.

ResponseColumn

Specify the name of the ResponseTable column that contains the response variable.

IDColumns

Specify the names of the ResponseTable and AttributeTable columns that specify the ID of the instance.

CategoricalEncoding

[Optional with CategoricalAttributeTable, disallowed otherwise.] Specify algorithm for encoding categorical columns:

Option	Description
GrayCode	Recommended when accuracy is critical. Depending on available memory, out-of-memory errors can occur if a categorical column has more than about 20 unique levels, even with a small data set.
Hashing	Optimizes calculation speed and minimizes memory use, possibly decreasing accuracy.

Default: 'GrayCode'

SplitsValueColumn

[Optional] If you specify SplitsTable, this syntax element specifies the name of the column that contains the split value.

Default: splits_valcol if UseApproximateSplits is 'true', attribute_value_column (the AttributeValueColumn syntax element value) otherwise

ApproxSplits

[Optional] Specify whether to use approximate percentiles (true) or exact percentiles (false). Internally, the function uses percentile values as split values.

For deterministic results, do all of the following:

• Specify ApproxSplits ('false').
• Use the UniqueID Syntax Element.
• For IdColumns, specify a single id_column that has a unique identifier for each observation.

Default: 'true' (nondeterministic)

NumSplits

[Optional] Specify the number of splits to consider for each variable. The num_splits_to_consider must be an INTEGER. If ApproxSplits is true, num_splits_to_consider must be greater than 1; otherwise, it must be greater than 0. The function does not consider all possible splits for all attributes.

Default: 10

MinNodeSize

[Optional] Specify the decision tree stopping criterion and the minimum size of any particular node within each decision tree.

Default: 100

MaxDepth

Specify a decision tree stopping criterion. If the tree reaches a depth past this value, the algorithm stops looking for splits. Decision trees can grow up to (2(max_depth+1) - 1) nodes. This stopping criteria has the greatest effect on function performance.

Default: 5

Maximum: 60

Weighted

[Optional] Specify whether to build a weighted decision tree. If you specify 'true', you must also specify the WeightColumn syntax element.

Default: 'false'

WeightColumn

[Required with Weighted, otherwise optional.] Specify the name of the response table column that contains the weights of the attribute values.

SplitMeasure

[Optional] Specify the impurity measurement to use while constructing the decision tree. If the tree is weighted, this value cannot be 'chisquare'.

Default: 'gini'

OutputProb

[Optional] Specify whether to output the probability distributions for each node.

Specify 'true' to use the OutputProb syntax element of the DecisionTreePredict_MLE function.

Default: 'false'

ResponseProbDistType

[Optional] Specifies the probability distribution type.

Default: 'laplace'