1.0 - 8.00 - CoxHazardRatio Example 4: RefFeatureColumns, Partition by ID - Teradata Vantage

Teradata® Vantage Machine Learning Engine Analytic Function Reference

Product
Teradata Vantage
Release Number
1.0
8.00
Release Date
May 2019
Content Type
Programming Reference
Publication ID
B700-4003-098K
Language
English (United States)

This example compares each of the four new patients in the table lc_new_predictors with each of the attribute reference values provided in the table lc_new_reference, partitioning by id and calculating the hazard ratio only when the patient id matches the reference id.

Input

SQL Call

SELECT * FROM CoxHazardRatio (
  ON lungcancer_coef AS cox_coef_model DIMENSION
  ON lc_new_predictors AS predicts PARTITION BY id
  ON lc_new_reference AS refs PARTITION BY id
  USING
  PredictFeatureNames ('trt','celltype','karno','diagtime','age','prior')
  PredictFeatureColumns ('trt','celltype','karno','diagtime','age','prior')
  RefFeatureColumns ('trt','celltype','karno','diagtime','age','prior')
  Accumulate ('id','name')
) AS dt ORDER BY 1, 2, 3, 4, 5, 6, 7, 8;

Output

Because the input is partitioned, the output table has only 4 rows. The attributes of patient Steffi are similar to those of reference id 4, so her hazard ratio is very close to 1.0 (0.97).

id name trt celltype karno diagtime age prior trt_ref celltype_ref karno_ref diagtime_ref age_ref prior_ref hazardratio
1 John standard squamous 30 4 63 yes   standard squamous 58 12 60 2.44014910130533
2 James standard large 80 12 41 no   standard smallcell 54 8 58 0.311881511740527
3 Stella test smallcell 70 3 72 no   test smallcell 52 12 61 0.502993022637894
4 Steffi test adeno 60 5 63 yes   test adeno 60 5 60 0.974218736747828