Input - Aster Analytics

This example uses the patient_pca_input data set and the variables age, bmi, bloodpressure, glucose, cigarettes, insulin, and hdl to predict the number of strokes for a patient.

The following statement creates the table pca_health_ev2 from a query that conducts a principal component analysis on the other seven variables:

CREATE TABLE pca_health_ev2 DISTRIBUTE BY REPLICATION AS
  SELECT * FROM pca_reduce (
    ON pca_map (
      ON pca_scaled
      TargetColumns ('age', 'bmi', 'bloodpressure', 'glucose',
                     'cigarettes', 'insulin', 'hdl')
    ) PARTITION BY 1
  ) ORDER BY component_rank;

The preceding query uses the same process as query in the PCA example.

The following query returns the table pca_health_ev2:

SELECT * FROM pca_health_ev2;

Because strokes is the response variable, pca_health_ev2 has no strokes column; otherwise, pca_health_ev2 is the same as the output table of the PCA example, pca_health_ev_scaled.