This topic provides details about the Teradata Scoring SDK functions supported by Scorer. Most details are the same as those for the corresponding SQL function; consult the section for the equivalent SQL function for a detailed description. The javadoc for Scorer provides information about how to run a scorer with a specific Teradata Scoring SDK function.
|Model Format||Model and its type accepted by the function. The model is incorporated inside AML the file.
Related syntax elements of AMLGenerator:
|Request Definition||Format for the input query to score against the Teradata Scoring SDK function, if different from the corresponding SQL function. In the context of a SQL function, this format corresponds to the input table schema for the test (query) data set. In most functions, this format is the same as the input used to train the model.
The format is supplied using the AMLGenerator syntax elements RequestColNames and RequestColTypes (to incorporate into the AML file).
Scorer creates an instance of a request object based on the format, which can be used to populate request data for scoring.
|Parameters||Parameters that the Teradata Scoring SDK function uses for scoring. These parameters correspond to the syntax elements of the corresponding SQL (predictor) function. This section also lists any syntax elements that the Teradata Scoring SDK function does not support.
These parameters can be supplied using AMLGenerator syntax elements RequestArgName1, RequestArgVal1, ReqestArgName2, RequestArgVal2, and so on (to incorporate into the AML file).
Because parameters can change frequently across queries for the same Teradata Scoring SDK function, parameters can also be supplied (with modifications) directly inside Scorer. See the Scorer javadoc for further information.
|[Optional] Additional Notes||Additional details about the function.|