Costs of vocabulary mapping

1
Costs of vocabulary mapping

• Leonard Will

http//www.willpowerinfo.co.uk
2
Different kinds of subject vocabularies

• Classification schemes
• Subject headings
• Thesauri
• Free text searching
• Terms and phrases extracted from text by computer

3
Ways of mapping

4
Input and output mapping

Users local thesaurus
User
Users free text terms
Local preferred terms
Mapping service thesaurus
Information providers preferred terms
Information providers thesaurus
Information providers database
Information providers preferred terms (perhaps
expanded)
5
Output mapping via remote thesaurus
Possible off-line consultation

Users local thesaurus
User
Users free text terms
Possible off-line term information
Mapping service thesaurus
Information providers preferred terms
Information providers thesaurus
Information providers database
Information providers preferred terms (perhaps
expanded)
6
Output mapping direct to remote database
Possible off-line consultation

Users local thesaurus
User
Users free text terms
Possible off-line term information
Mapping service thesaurus
Off-line provision of term information
Information providers preferred terms
Information providers thesaurus
Information providers database
7
Other projects and tools

• CARMEN
• RENARDUS
• AQUARELLE
• SIS-TMS
• UMLS
• GenThes
• CERES
• Knowledgecite Library

8
Personal experience

• Editing terms into thesaurus structure consistent
  with AAT
• 25 terms per hour 150 terms per day
• 4500 terms take 30 days
• Assigning Dewey numbers to UNESCO terms
• 15 terms per hour 90 terms per day
• 4500 terms take 50 days
• Note AAT has about 125,000 terms
• At 90 terms per day would need 1400 days 6.3
  years

9
English Heritage estimator for thesaurus
construction

• Based on creating 10-20 simple terms per day
• Complex terms 2-8 per day
• At 10 terms per day,
• 4,500 terms take 450 person-days
• just over 2 person-years

10
Factors affecting calculation

• Number of terms
• Number of uses
• Candidate terms per year
• Number of external terms mapped per year
• Number of licenses

11
Start up requirements (EH)

• User assessment / market testing 3 days
• Introduction 10 days
• Peer review 5-15 days
• Initial documentation 2 days
• Promotion 3 days
• Audit of existing usage 0.1 day per 500 uses
• Research (e.g. reconciliation of names)
  10 require 1 day per 5 terms
• Training 1 day per 5 licences

12
Annual maintenance tasks

• Candidate term evaluation
• 1 day per 5 terms received
• Mapping of existing terminology
• 1 day per 50 terms received
• Tracking and version control
• 1 day per 1000 terms
• License management
• 0.5 days per license
• (c) English Heritage 2001 .

13
One to one mapping

• Abandoned children ? 305.906945 Abandoned
  children
• Abbreviations ? 401.48 Abbreviations
• Ability ? 153.9 Intelligence and aptitudes
• Ability grouping ? 371.254 Homogeneous grouping

14
One to many mapping

• Abortion ?
• 179.76 Abortion (ethics)
• 294.356976 Abortion (ethics - religion -
  Buddhism)
• 304.667 Abortion (demographic effects)
• 342.084 Abortion (law and comprehensive works)
• 342.085 Abortion (rights of fetuses)
• 342.0878 Abortion (rights of women)
• 344.04192 Abortion (medical law)
• 363.46 Abortion (social problems)
• 363.96 Abortion (birth control)
• 364.185 Abortion (criminal offences)
• 615.766 Abortion (drugs causing)
• 618.392 Abortion (spontaneous)
• 618.88 Abortion (surgical)