Managing binning options for numeric predictors

Generate bins over which you want to distribute numeric predictors. Binning allows you to group cases into bins of equal volume or width. The bins of equal width are simpler to manage and easy to understand. The bins of equal volume scale data well, handle skewness better, but are difficult to use with categorical data.

  1. In the Data analysis step, click a numeric predictor.

  2. Click the Binning tab.

  3. In the Minimum value field, specify the threshold value of the lowest value bin.

    Values lower than the minimum value belong to the lowest bin.

  4. In the Maximum value field, specify the threshold value of the highest value bin.

    Values higher than the maximum value belong to the highest bin.

  5. In the Number of intervals field, specify the number of intervals over which the values of predictors will be analyzed.

  6. Select whether you want to create intervals with equal volume or width.

  7. Click Add special value to add a value that you want to exclude from the bins. Typically, special values represent error codes.

  8. Click Apply.

  9. Click the Graphical view tab or Tabular view tab to read the data in the graph or table.
  10. If you change the field properties, you can select the Use as default settings option to override the default project settings.

  11. Click Submit.