Output Table Schema
The table has a set of predictions for each test point.
| Column | Data Type | Description | ||||||
|---|---|---|---|---|---|---|---|---|
| accumulate_column | Same as in input table | [Column appears once for each specified accumulate_column.] Column copied from input table. | ||||||
| id_column | Same as in input table | Column copied from input table. Unique row identifier. | ||||||
| prediction | VARCHAR | Predicted test point value or predicted class, determined by model. | ||||||
| confidence_lower | DOUBLE PRECISION | [Column appears only with OutputProb ('false').] Lower bound of confidence interval. For classification trees: confidence_lower = confidence_upper = t / T Where:
For regression trees: confidence_lower = confidence_upper = prediction ± 1.96 * sd(prediction) / sqrt(T) |
||||||
| confidence_upper | DOUBLE PRECISION | [Column appears only with OutputProb ('false').] Upper bound of confidence interval. For how function calculates probability, see description of confidence_lower. | ||||||
| tree_num | VARCHAR | [Column appears only with Detailed ('true').] Either the concatenation of task_index and tree_num from Model, to show which tree generated prediction, or 'final' to show overall prediction. | ||||||
| prob | DOUBLE PRECISION | [Column appears only with OutputProb ('true') and without Responses syntax element.] Probability that observation belongs to class prediction, calculated as follows: t / T Where t and T are as described in confidence_lower. |
||||||
| prob_response | DOUBLE PRECISION | [Column appears only with Responses syntax element, and appears once for each specified response.] Probability that observation belongs to category response, calculated as follows: t / T Where:
|