Skip to main content
Version: Next

Fit Data Preparation

Good data preparation is the most important factor for stable fitting outcomes.

Preparation Checklist​

  1. Ensure variable units are consistent.
  2. Remove obvious outliers before fitting.
  3. Keep enough samples to identify unknown coefficients.
  4. Keep variable ranges representative of the target operating region.

Table Preparation Patterns​

  1. Forward calculation:
    • Fill variable columns only.
  2. Coefficient fitting:
    • Fill variable columns and actual-value column.
  3. Target inversion:
    • Fill actual-value column; optionally provide variable initial guesses.

Quality Gates​

  1. No empty required cells.
  2. No non-numeric values in numeric fields.
  3. No duplicated rows that add no information.
  4. No unit mismatch across rows.

Suggested Sampling Strategy​

  1. Cover both low and high ends of operating ranges.
  2. Avoid all samples clustered in a narrow range.
  3. Add independent verification rows not used for fitting.
  1. Fit Basic Operations
  2. Fit Result Validation