RandomSample Example 2: KMeans++ Sampling - Teradata Vantage

This example uses KMeans++ sampling with the Manhattan distance metric, and treats the numeric variables cyl, gear, and carb as categorical variables (and the categorical variables vs and am). The category weights are assigned in the order that the columns appear in the input table: 1000 to cyl, 10 to vs, 100 to am, 100 to gear, and 100 to carb.

Input

• InputTable: fs_input, as in RandomSample Example 1: Basic Sampling (Weighted)

SQL Call

SELECT * FROM RandomSample (
  ON fs_input AS InputTable
  USING
  NumSample (10)
  SamplingMode ('kmeans++')
  InputColumns ('mpg:carb')
  CategoryWeights (1000, 10, 100, 100, 100)
  AsCategories ('cyl', 'gear', 'carb')
  Distance ('manhattan')
  Seed (1)
  SeedColumn ('model')
) AS dt ORDER BY 1, 2, 3;

Output