Who should use the Model Evaluation workflow?
Teams or solo builders working on development tasks who want a repeatable process instead of one-off tool experiments.
Execution Map
3-step pipeline — follow each step in order
1Setup · Step 1 of 3
Open task pageSet up model for evaluation
Deploy the AI model to a test environment using MathWorks MATLAB AI to prepare for performance evaluation. This ensures the model is accessible for testing with validation data.
Why this matters
Deploying the model is necessary to create a controlled environment where evaluation metrics can be accurately measured without interference.
What you get
The model is deployed and ready to accept inputs for evaluation, enabling the next step of performance assessment.
2Input · Step 2 of 3
Open task pageRun model evaluation
Execute the model evaluation using Forefront AI to compute key metrics such as accuracy, precision, recall, and F1-score on the validation dataset. This provides a quantitative assessment of model performance.
Why this matters
This is the core step where the model’s performance is measured, providing the primary deliverable of the workflow.
What you get
A comprehensive evaluation report with performance metrics is generated, highlighting the model’s strengths and weaknesses.
More optionsCompare
3Deliver · Step 3 of 3
Open task pageMonitor ongoing performance
Use SAS Viya to set up continuous monitoring of the model’s performance over time, detecting drift or degradation. This ensures the model remains reliable after deployment.
Why this matters
Monitoring is critical to catch performance issues in production, allowing for timely retraining or adjustments.
What you get
A monitoring dashboard is established, providing real-time alerts and trends for model performance.
More optionsCompare