How does “Refine your results” work?

The difference in the Content Mining (CM) results is given by the difference in the results returned by the sources.

Using “Full Text” limiters limits the number and the ‘quality’ (with regard to the CM) of the results returned by the search sources. The configuration of the module and the different results returned by the different searches are factors that limit the number of terms in the “Refine your results” window. We actually don’t display any term unless its threshold weight is above 3, which usually means that the term must be found at least 2-3 times in the returned records.

Of course the limit can be lowered to 1 so that the CM always returns something, no matter the search, but the results will not always be meaningful with respect to the search query. It will also waste quite a lot of heap memory since we will have to keep more terms into the main memory until they are being sent to the browser to be displayed. Having many CM terms is less important than having only the relevant terms even if it means that there is only one relevant term. The CM process must be focused on quality and not quantity.