With a growing number of factoids we might be able to create a statement type suggestion from text similarity: e.g. by overlap in terms, by tf/idf similarity, or by embedding similarity (check vector index functionalities of solr?) to reduce inconsistent modelling.
With a growing number of factoids we might be able to create a statement type suggestion from text similarity: e.g. by overlap in terms, by tf/idf similarity, or by embedding similarity (check vector index functionalities of solr?) to reduce inconsistent modelling.