Amazon SageMaker Canvas supports training ML models with different objective metrics
Amazon SageMaker Canvas now provides the ability to train machine learning (ML) models with different objective metrics, allowing you to gain a more comprehensive understanding on the model's strengths and weaknesses. SageMaker Canvas is a visual interface that enables business analysts and citizen data scientists to generate accurate ML predictions on their own — without requiring any ML expertise or having to write a single line of code.
Amazon SageMaker Canvas now provides the ability to train machine learning (ML) models with different objective metrics, allowing you to gain a more comprehensive understanding on the model's strengths and weaknesses. SageMaker Canvas is a visual interface that enables business analysts and citizen data scientists to generate accurate ML predictions on their own — without requiring any ML expertise or having to write a single line of code.
By default SageMaker Canvas selects the most suitable objective metric for each problem type. However, by training ML models with different objective metrics, you can enhance their robustness and generalization capabilities. Optimizing for a single metric may lead to overfitting or bias towards the training data. Different metrics often involve trade-offs. For instance, optimizing for precision may result in lower recall, and vice versa. By training models with different objective metrics, you can assess these trade-offs and choose the best compromise that aligns with your specific requirements. Until now, SageMaker Canvas only supported a single default objective metric for each problem type. Starting today, you can select an objective metric from the list of supported metrics and optimize your ML models accordingly.
The ability to train ML models with different objective metrics in Amazon SageMaker Canvas is now available in all AWS regions where SageMaker Canvas is supported. To learn more, refer to the SageMaker Canvas product documentation.