This is like something that a Google search displays towards the bottom of the page: "Related Searches."
"Related" to be measured by parsing the lookup table of Topics. Could be given a weighted score for relatedness to other topics, though this is large.
Maybe the topics could be set up in a taxonomy that contains a huge number of non-hierarchical "related terms." There will also be some narrower and broader terms in our corpus of subject terms, including the curated topic list. The system could select from a set of "related terms" randomly or using some metric, plus one narrower or broader term, if there is one.
Display a closely RELATED topic or two
This is like something that a Google search displays towards the bottom of the page: "Related Searches."
"Related" to be measured by parsing the lookup table of Topics. Could be given a weighted score for relatedness to other topics, though this is large.
Maybe the topics could be set up in a taxonomy that contains a huge number of non-hierarchical "related terms." There will also be some narrower and broader terms in our corpus of subject terms, including the curated topic list. The system could select from a set of "related terms" randomly or using some metric, plus one narrower or broader term, if there is one.