Columns in Data Mining Models
Columns define the attributes that the mining algorithm uses to discover patterns. Each column has a set of properties that control how the algorithm treats the data.
| Column Name | Data Type | Role | Continuous? | Key Column? | Missing Value Treatment |
|---|---|---|---|---|---|
| CustomerID | int | Key | No | Yes | Ignore |
| Age | int | Predictor | Yes | No | Average |
| Gender | nvarchar(10) | Predictor | No | No | Mode |
| Income | money | Predictor | Yes | No | Average |
| PurchaseHistory | nvarchar(max) | Predictor | No | No | Ignore |
| Response | int | Target | No | No | Ignore |
MDX Example: Selecting Column Information
SELECT
{[Measures].[Column Cardinality],
[Measures].[Column Avg Length]}
ON COLUMNS,
{[Column].[CustomerID],
[Column].[Age],
[Column].[Gender],
[Column].[Income],
[Column].[PurchaseHistory],
[Column].[Response]}
ON ROWS
FROM [MiningModel];
This query retrieves metadata for each column in the MiningModel object.