A Tale of Two Taxonomies

Taxonomies occupied a double slot ("Taxonomy Tales") on the program on Thursday afternoon. Jennifer Evert, Taxonomy Product Manager at LexisNexis and Marjorie Hlava, President of Access Innovations each discussed their work on taxonomies.

LexisNexis

At LexisNexis, taxonomies and indexing are seen as a key strategic focus to help users find answers quickly and efficiently. They are used to analyze answer sets and suggest search terms, classify sources, and launch new searches. The power of taxonomies lies in their ability to manage large amounts of information, provide a common terminology, create metadata to enhance the search process, and help users find answers to make more informed decisions.

The LexisNexis taxonomy has more than 3500 subject or industry terms, 330,000 company names, and 900 geographic area terms. Many of the terms were created using rules developed by the indexers. The rules base includes rules for weighting, location, frequency, and variety of terms. Expert indexers research, test, and maintain the rule sets, and the entire archive is re-indexed quarterly. Automated indexing helps develop terms, but human editorial expertise is still needed. The coverage is currently being extended from English to other languages; German and French are the first non-English languages to be included. LexisNexis has developed parallel interfaces: one for sophisticated searchers that takes full advantage of the taxonomies and thesaurus metadata to build queries, and one for casual users that provides simple search aids and analysis of answer sets to help them fine-tune their searches.

Access Innovations

Access maintains the thesaurus of the National Information Center for Educational Media (NICEM). Developers have access to a tool that provides them with various views of the thesaurus: alphabetic, hierarchical, and term records. Extensive use is made of machine-aided indexing (MAI), which suggests terms, tracks suggestions and editor’s choices, and enables rule changes. However, MAI is only an aid and prompt; the editor makes the decisions about term inclusion. The editors can also write rules based on term selection criteria.

Requirements for indexing tools include: suggesting terms that are valid, correctly formatted, and conceptually appropriate, but not suggesting terms not meeting these criteria. Good indexing tools will lead to faster and more consistent production by providing memory prompts for forgotten terms, indexing all relevant concepts deeply and specifically, and integrating the thesaurus with the database management system.

Searchers find info 50% faster using browsable categories instead of list returned from a full text search. They prefer browsable category search terms. Integrated thesaurus tools provide cross-checks, validation, feedback, error prevention, interconnection, cooperation, coordination, and seamless integration with the database.

With all these benefits, there is little wonder why thesauri and taxonomies are such popular subjects today.

Don Hawkins
Columnist, Information Today

Comments are closed.