Last month, new natural language processing and machine learning tools were released in beta format in HeinOnline’s Law Journal Library. This blog post described the new More Like This tool, as well as new topic and entity application, in detail. This month, additional features using these concepts are now available.
Topics and entities are now available within document metadata fields and search facets. Entities include location, person, and organization. For instance, search for “Supreme Court” AND gerrymander* and sort results by number of times cited by articles. In the facets, topic includes constitutional law and politics; location includes United States and New York; person includes Brown and Scalia; and organization includes Supreme Court and Congress.
NOTE: Search facets have been reordered based on usage. Due to the rich metadata indexing available throughout HeinOnline, these facets are also now expandable and collapsible in order to keep the left viewing pane clean and easy to use.
Practical Application Example
For HeinOnline researchers, the addition of these topics and entities to document metadata supplies a new benefit: search results are enhanced to include even more relevant articles.
Run the same search, but insert text: before each phrase. This circumvents searching the metadata, instead searching only document text. The first result is missing, because the exact phrase “worker’s compensation” does not appear in the article’s text, although it concerns the subject. In fact, many of the articles in the results of this second search actually cite the missing article!
More Like This, Topics, and Entities are currently all in beta format in the Law Journal Library, with plans to expand to other HeinOnline content in the near future. This gives users the opportunity to utilize the new features and provide feedback on their usability and on areas of potential improvement.