Subject cataloguing:

Subject cataloguing:

faster, better, cheaper

Wan Wong & Alison Dellit

National Library of Australia

8560 B.C.Cultivation of emmer, barley, and other wild grasses begins on the steppes of Central Asia, leading to the development of specialized occupations, such as cereal catalogers

“Great moments in the history of technical services” (http://sun3.lib.uci.edu/~murrizol/ts_history/tshist.htm)

Subject cataloguing

• Time consuming but essential

• Based on LCSH

• Many places to look

1. Brings together different data

• LCSH (including subdivisions)

• Australian extension to LCSH

• NLA specific subject headings

• ANBD

• Dewey

2. Searches whole authority record

• Keyword searching

• Main heading, references & scope notes

• Relevance-ranked results

3. Searches ANBD

• Number of records

• From NLA / from LC

• Date last used

4. Suggests DDC

• Dewey data– Numbers– Hierarchy– Mapping methods

• DDCs from ANBD, particularly recent NLA records

5. Interacts with ILMS

• Activated from ILMS

• Exports part or whole strings back with correct tags & subfields

• DDC as well

Mock-up of the tool

Where to from here?

Automated Metadata Extraction

– Based on semantic analysis– Reasonably good at working out

creator, resource type, dates, but not good at subject analysis.

Automated Classification

– Based on analysing patterns in articles that have already been correctly classified.

– Has good success rate with short full-text articles, or articles all in the same field.

– Recommind is an example.

Our role into the future

“To each book, its reader”

- Ranganathan

For more information

[email protected]

[email protected]