Our goal in this chapter was to make predictions using trained predictive or machine learning models. In the first section of this chapter, we learned how to select the right version of the predictive model. We used the two models of versioning described in Chapter 9, Predictive Model Training and Evaluation. In the second section, we learned that we need to provide data and parameters from the T-SQL part of the SQL Server to external scripts. We looked at how to pass parameters to the external script executed by the sp_execute_external_script stored procedure. We also explored how to consume results from the stored procedure.
In the third section, we discussed the use of the PREDICT keyword as an alternative to the predictions calculated by external scripts. In this section, we also considered the limits of the PREDICT keyword. In the final section, we joined our work...