Topicbased approaches to searching

1
Topic-based approaches to searching
Judith Pearce and Basil Dewhurst National Library
of Australia
2
What Ill cover

• The ongoing need for controlled vocabularies and
  authority control
• How they are currently being deployed in resource
  discovery
• Doing more with authority files through a
  topic-based approach to searching
• People Australia as exemplar

3
Assertion

• Findability is improved through the use of
  controlled vocabularies

4
Why

• A single word may have multiple meanings
• Different words may mean the same thing
• A concept may be a subset of a broader concept or
  related to another concept
• People may have many names
• The same name may be used by different people

5
Types of controlled vocabularies

• Simple enumerated lists
• Taxonomies
• Synonym rings
• Thesauri
• Rules-based

6
IFLA conceptual models

• Functional Requirements for Bibliographic Records
  (FRBR)
• Functional Requirements for Authority Data (FRAD)
• Functional Requirements for Subject Authority
  Records (FRSAR)

7
Bibliographic entities
Person
Concept
Work
Family
Object
Expression
Event
Manifestation
Corporate body
Place
Item
Names
Identifiers
Controlled access point
8
International thesaurus standards

• ISO 27881985 Guidelines for the establishment
  and development of monolingual thesauri (BS5723
  1987)
• ISO 5964 1985 Guidelines for the establishment
  and development of multilingual thesauri (BS
  6723 1985)

9
New work item proposal

• Revise the two ISO standards based on BS
  87232005 Structured vocabularies for information
  retrieval
• Definitions, symbols and abbreviations
• Thesauri
• Vocabularies other than thesauri
• Interoperability between vocabularies
• Exchange formats and protocols for
  interoperability

10
Other work

• ANSI/NISO Z39.19 2003 Guidelines for the
  Construction, Format, and Management of
  Monolingual Thesauri
• Freely available
• Single part
• Readable
• Addresses interoperability issues
• Touches on multilingual issues

11
Schemas

• Authority data
• International Standard Archival Authority Record
  for Corporate Bodies, Persons, and Families
  (ISAAR)
• MARC format for authority data
• Metadata Authority Description Schema (MADS)

• Thesauri
• Z39.50 ZTHES
• Simple Knowledge Organisation System (SKOS)
• Parties
• Encoded Archival Context (EAC)
• DC Agents
• International Standard Name Identifier (ISNI)

12
Deployment (traditional)

• Keyword searching
• part of resource description
• Limits and filters
• Coded values not directly searchable
• Fielded searching
• two-step process to exploit non-preferred forms
• Subject hierarchies
• Authority files
• Mainly a librarians tool
• Need for user training and intermediation

13
Deployment (new generation)

• Google-like search
• Relevance ranking
• Faceted clustering
• Form (books, pictures, maps)
• Names
• Subjects
• Dates
• Audience

14
Outstanding issues

• What the user keys in may not match controlled
  access points
• Limitations of faceted clustering
• How many topics to display
• Selecting a facet limits the search
• Increases precision but not recall
• Topic attributes not exploited in the simple
  keyword search

15
Topic-based searching
16
Another way of looking at FRBR
Topic
Resource
Party
Work
Concept
Expression
Event
Manifestation
Place
Item / Object
Resource
17
People Australia

• An online service that
• Takes a topic based approach to resource
  discovery
• Provides access to records about people and
  organisations
• Allows users to find and get resources, by/about
  a person or organisation
• Interoperates with contributors services

18
Precursors

• Bright Sparcs
• Australian Womens Archive Project
• Australian Trade Union Archive
• Dictionary of National Biography
• Dictionary of Australian Artists Online
• Archive manuscript collections
• Australia Dancing
• Music Australia

19
Content

• Australian Name Authority File
• 200,000 Australian names
• Some contextual information
• Contributor data
• 24 archives, museums and libraries
• Over 47 services
• Rich sources of information
• Resources by or about Australian people in
  Australian collections

20
Architecture

• Harvesting contribution model
• Records clustered under a single identifier
• Harvested records fully searchable
• Hub or lens pages for each person
• Links to information by or about the person
• Deep linking to contributor services
• Both people and organisations

21
Example

• Sir Douglas Mawson was a geologist and explorer
  famous for leading the Australasian Antarctic
  Expedition 1911 1914
• A People Australia page for Mawson would provide
• Biographical information from contributor records
• Bibliographic records from Libraries Australia
• Image records from Picture Australia
• Newspaper articles from the new Newspapers
  Discovery Service
• Related scholarship from the ARROW Discovery
  Service
• Museum objects
• Archival resources

22
Data Schema

• Based on the Encoded Archival Context (EAC)
  standard
• Alternatives considered
• MARC/MARCXML/MADS
• Interparty
• Dublin Core Agent
• Music Australia Party Schema (MAPS)

23
System interfaces

• Harvest
• OAI-PMH
• Search
• SRU, OpenSearch, Z39.50 (via SRU gateway)
• Augment
• SRU Update, OAI-PMH
• Syndicate
• OAI-PMH, Coins, RSS, etc
• Resolve
• Info URI (to lens page)
• Search protocols or openURL (to metadata)

24
OAI-PMH

• Harvesting
• Metadata formats
• EAC
• Dublin Core
• MARCXML
• Syndicating
• EAC
• Dublin Core

25
Project progress

• Work is proceeding on building the Search and
  Harvest services
• Search interfaces available in second qtr 2008
• Will be making prototypes widely available for
  user testing and feedback
• Will create a public website to provide
• Information about the project
• Regular news
• Technical information

26
Topic-based approaches to searching

• Judith Pearce
• Phone [phone]
• Email jpearce_at_nla.gov.au
• Basil Dewhurst
• Phone [phone]
• Email bdewhurst_at_nla.gov.au