How do we describe something?

1
How do we describe something?

• What something is about?
• What the content of an object is about?
• Different methods (Wilson, 1968)
• counting terms (objective method)
• complete description/summarization
• unifying thought(s)
• What stands out (main points)
• Challenges
• Non-text

2
Languages for aboutness

• Indexing languages
• Terminological tools
• Thesauri (CV controlled vocabulary)
• Subject headings lists (CV)
• Authority files for named entities (people,
  places, structures, organizations)
• Classification
• Keyword lists
• Natural language systems (broad interpretation)

3
Aboutness How to do it!

• Read the document Intellectual reading
• look for key features
• many indexers mark up the items
• rarely have time to read the whole document
• Determine aboutness Conceptual analysis
• Translate aboutness into the vocabulary or scheme
  you are using
• In general Subject headings 1-3 headings
• Descriptors, 5-8 descriptors
• Classification 1 notation (should it only be
  one!?).

4
Features of indexing languages

• With the exception of a few general domain tools,
  they are generally domain specific.
• MeSH
• NASA Thesaurus
• Astronomy Thesaurus
• ERIC thesaurus
• http//www.darmstadt.gmd.de/lutes/thesoecd.html
• Concepts (or concept representations) are
  arranged in a discernable order?

5
Language schema designs

• Classified--grouping
• Hierarchies and facets
• MeSH Browser
• http//www.nlm.nih.gov/mesh/MBrowser.html
• Art and Architecture (Getty AAT)
• http//www.getty.edu/research/conducting_researc
  h/vocabularies/aat/
• Alphabetical -- horizontal
• Verbal/Alphabetical (ordering/filing challenges)

6
Controlled Vocabulary

• Why do we have a controlled vocabulary?
• Three of you independently identify a new human
  gene, and each separately name it different
  things.
• How do we handle references/resolving/utilizing
  this concept which has different names. Let
  alone, across languages?!

7
Controlled Vocabulary

• A list or a database of subject terms in which
  each concept has a preferred terms or phrase that
  will be used to represent it in the retrieval
  tool the terms not used have references
  (syndetic structure), and often scope notes.
  Their can be aliases for preferred terms (so the
  all three of your gene names get recorded and are
  matchable to the preferred term).

8
Example

• For gene names, there is an authority, HUGO Gene
  Nomenclature Committee that designates an
  official curated name for gene.
• During the research process however, there may
  have been multiple initial names.

9
More Examples

• Most processs however, do NOT have standardized
  naming.
• For instance genetic conditions are not named in
  one standard way. Doctors treating patients
  often propose the first name, but often expert
  working groups later revise to more appropriate
  name.

10
Contd

• The basic genetic or biochemical defect that
  causes the condition (for example, alpha-1
  antitrypsin deficiency)
• One or more major signs or symptoms of the
  disorder (for example, hypermanganesemia with
  dystonia, polycythemia, and cirrhosis)
• The parts of the body affected by the condition
  (for example, craniofacial-deafness-hand
  syndrome)
• The name of a physician or researcher, often the
  first person to describe the disorder (for
  example,Marfan syndrome, which was named after
  Dr. Antoine Bernard-Jean Marfan)
• A geographic area (for example, familial
  Mediterranean fever, which occurs mainly in
  populations bordering the Mediterranean Sea) or
• The name of a patient or family with the
  condition (for example, amyotrophic lateral
  sclerosis, which is also called Lou Gehrig
  disease after the famous baseball player who had
  the condition).

11
Thesaurus (structured thesaurus)

• Lexical semantic relationships
• Composed of indexing terms/descriptors
• Descriptors representations of concepts
• Concepts Units of meaning (Svenonius)

12
Thesaurus

• Preferred terms
• Non-preferred terms
• Semantic relations between terms
• How to apply terms (guidelines, rules)
• Scope notes
• Adding terms (How to produce terms that are not
  listed explicitly in the thesaurus)

13
Preferred Terms

• Control form of the term
• Spelling, grammatical form
• Theatre / Theater
• MLA / Modern language association
• Choose preferred term between synonyms
• Brain cancer or Brain Neoplasms?

14
Common thesaural identifiers

• SN Scope Note
• Instruction, e.g. dont invert phrases
• USE Use (another term in preference to this one)
• UF Used For
• BT Broader Term
• NT Narrower Term
• RT Related Term

15
Semantic Relationships

• Hierarchy
• Equivalence
• Association

16
Hierarchies of Meaning
Beer Glass

White wine glass

Glass

Wine Glass

Red wine glass

From Controlled Vocabularies/ Paul Miller
Interoperability Focus UKOLN
17
Hierarchy

• Level of generality both preferred terms
• BT (broader term)
• Robins BT Birds
• NT (narrower term)
• Birds NT Robins
• Inheritance, very specific rules

18
Equivalence

• When two or more terms represent the same concept
• One is the preferred term (descriptor), where all
  the information is collected
• The other is the non-preferred and helps the user
  to find the appropriate term

19
Equivalence

• Non-preferred term USE Preferred term
• Nuclear Power USE Nuclear Energy
• Periodicals USE Serials
• Preferred term UF (used for) Non-preferred term
• Nuclear Energy UF Nuclear Power
• Serials UF Periodicals

20
Association

• One preferred term is related to another
  preferred term
• Non-hierarchical
• See also function
• In any large thesaurus, a significant umber of
  terms will mean similar things or cover related
  areas, without necessarily being synonyms or
  fitting into a defined hierarchy

21
Association

• Related Terms (RT) can be used to show these
  links within the thesaurus
• Bed RT Bedding
• Paint Brushes RT Painting
• Vandalism RT Hostility
• Programming RT Software