Fit()#

Description#

Fit() computes and stores the minimum and maximum values for the specified columns from the provided table.

The computed statistics are stored internally and later used during Transform to perform range-based scaling.

Fit() overwrites any previously stored statistics for the same columns.

Signature#

func (s *RangeScaler) Fit(t *table.Table, columns ...string) error

Parameters#

  • t#

    *table.Table
    The input table used to compute column statistics.

  • columns#

    ...string
    Zero or more column names to be fitted.

    Only the specified columns are processed.

    Calling Fit() with zero column names results in a no-op. In this case, IsFitted() remains returning false.

Return Value#

  • error#

    Fit() returns an error if:

    • The provided table is nil
    • A specified column does not exist
    • A column contains no numeric values

    If all columns are fitted successfully, nil is returned.

Behavior#

  • Computes minimum and maximum values for each specified column
  • Considers only numeric values when computing statistics
  • Ignores non-numeric values during computation
  • Stores computed statistics internally
  • Appends fitted column names to the feature list
  • Overwrites previously fitted statistics for the same columns
  • Does not modify the input table
  • If no column names are provided, Fit() performs no operation and the scaler remains unfitted

Example Usage#

After creating a RangeScaler instance, call Fit() with a table and the columns to be scaled.

Once Fit() succeeds, the scaler is ready to be used with Transform().

When to Use#

Use Fit() when you need to:

  • Learn per-column minimum and maximum values from a reference dataset
  • Prepare a RangeScaler for consistent range-based normalization
  • Refit the scaler with updated or different training data

See Also#

Backward RangeScaler Transform() Forward