Refinements and Binning
Refinements, also known as Binning, is a new way of navigating through search collection results based on grouping subsets of the results matching a query into "refinement sets" or "bins."
The refinements engine requires only two inputs:
- The documents and content elements that are to be sorted into bins.
- The specification of how to create those bins.
Currently, Watson™ Explorer Engine supports binning at the search-collection level only. Most commonly, the binning will be generated by only one search collection, however it is possible to merge the binning output of multiple search engine collections, so that multiple data sources can be navigated with the same binning interface.
The data that is used to create the bins come from contents (contained in documents) that have been fast-indexed by the Watson Explorer Engine Search Engine. A list of contents to include in the fast index can be specified in the "Configuration" tab of a search engine collection under the subtab "Indexing".
The binning configuration can be specified in the "Configuration" tab of your search engine collection under the "Binning" subtab. This simple interface provides a way to add binning "sets" to the binning configuration, and then binning "ranges" and "trees" as well as additional sets as children to any existing sets.
When a binning configuration is specified for a search engine collection, a binning element will be included in the results returned from the search engine. This element will contain a set of bins which have been generated from the binning-sets specified in the configuration.
The bundled IBM display render-binning and render-sliderbin displays use list- and graphical bins (respectively), but you may want to customize those controls to make them easier for your end-users to understand, or customize them to the specific content of your collection.
For a hands-on tutorial of creating and using graphical and list bins with a search collection, see Tutorial: Using Refinements to Simplify Result Selection.