https://communities.sas.com/t5/SAS-Programming/Model-honest-assessment/m-p/488711#M127419 <P>The terms can vary a bit. But the idea is three datasets:</P> <P>&nbsp;</P> <OL> <LI>training - used to develop a model</LI> <LI>validation - used to calibrate model. If model doesn’t perform well, revisit model and use this data to validate</LI> <LI>test - used in final stages, to get final model accuracy &nbsp;numbers. Used to prevent over fitting and provide an 'honest assessment’. You would not retrain a model if this didn’t score well&nbsp;</LI> </OL> Tue, 21 Aug 2018 20:42:02 GMT Reeza 2018-08-21T20:42:02Z https://communities.sas.com/t5/SAS-Programming/Model-honest-assessment/m-p/488692#M127414 <P>Is the test data and the validation dataset the same?</P> Tue, 21 Aug 2018 19:26:02 GMT https://communities.sas.com/t5/SAS-Programming/Model-honest-assessment/m-p/488692#M127414 ajkalale 2018-08-21T19:26:02Z https://communities.sas.com/t5/SAS-Programming/Model-honest-assessment/m-p/488711#M127419 <P>The terms can vary a bit. But the idea is three datasets:</P> <P>&nbsp;</P> <OL> <LI>training - used to develop a model</LI> <LI>validation - used to calibrate model. If model doesn’t perform well, revisit model and use this data to validate</LI> <LI>test - used in final stages, to get final model accuracy &nbsp;numbers. Used to prevent over fitting and provide an 'honest assessment’. You would not retrain a model if this didn’t score well&nbsp;</LI> </OL> Tue, 21 Aug 2018 20:42:02 GMT https://communities.sas.com/t5/SAS-Programming/Model-honest-assessment/m-p/488711#M127419 Reeza 2018-08-21T20:42:02Z