Skip to main content
Version: Next

Fit Result Validation

Validate Fit results before writing them back or using them in downstream processes.

Validation Dimensions​

  1. Numerical convergence status.
  2. Residual magnitude and trend.
  3. Coefficient plausibility.
  4. Predicted-vs-actual consistency.

Step-by-Step Validation​

  1. Check operation status in result window.
  2. Review iteration count and termination reason.
  3. Check residual norm against your acceptance threshold.
  4. Compare predicted values with actual values on key rows.
  5. Verify solved variables or coefficients are within physical bounds.

Acceptance Example​

A practical acceptance rule can be:

  1. Converged status.
  2. Residual norm below predefined threshold.
  3. No coefficient at hard lower/upper bound unless expected.

Rejection Signals​

  1. Not converged with large residual.
  2. Coefficients pinned at bounds unexpectedly.
  3. Large prediction errors in critical operating region.
  1. Fit Data Preparation
  2. Fit Common Scenarios