FRBR: Algorithms and Applications

T. HickeyJ. TovesD. Vizine-Goetz

Online Compuer Library Center

CLA November 2004

Outline

Algorithms• FRBR work matching• Handling author-title variants

Hardware• Beowulf cluster

Applications• Bookmarklets• FictionFinder

Future directions

Working with Group 1 Entities

WEMI:Work

ExpressionManifestation

Item

Strict expression-level determination is hard• We primarily divide by language

Manifestation is easier• We use the WorldCat master record

Work Identification

Algorithm goals:• Efficient• Understandable• Controllable by catalogers• Uses existing WorldCat records

The Algorithm

A key is generated for each record Extract author, title

• Look up in LC name authority file• Added entry information as needed

Form a key from bibliographic record• Author, title, added entry information• These can be sorted, compared

Example

146 Smollett\1721 Expedition of Humphry Clinker

16 Smollett\1721 Expedition of Humphrey Clinker

8 Smollett\1721 Humphry Clinker

4 Smollett\1721 Humphrey Clinker

2 Smollett\1721 Expedition of Humphry Clinker

1 Smollett\1721 Calatoriile lui Humphrey Clinker

1 Smollet\1721 Expedition of Humphry Clinker

1 Smollett Humphry Klinkers Reisen

Example (with authorities)

156 Smollett\1721 Expedition of Humphry Clinker

16 Smollett\1721 Expedition of Humphrey Clinker

4 Smollett\1721 Humphrey Clinker

1 Smollett\1721 Calatoriile lui Humphrey Clinker

1 Smollet\1721 Expedition of Humphry Clinker

1 Smollett\1721 Humphry Klinkers Reisen

More Detail

Extract author names• Look up in authority file

• Currently only personal names• Subfields $abcdq

Extract title• Always use uniform titles if present• Look up author/short title (~$a)• Look up author/long title (~$abfgnp)• Prefer alternative title for non-English

Create key from author/title• Always do NACO normalization (has limitations)• Add information for uncontrolled title-main-entry

Authority Files Rule!

Authors Author/titles

Bring together variations Allow override in difficult cases

• Both splitting and joining groups• Especially important with xISBN matching

Especially important with non-English metadata

Limitations of the Authority File

What’s missing:• Many uniform titles• Many author variants• Many title variants• Language of heading

Partial solution• Create auxiliary files of mechanically generated matches

Results of FRBR Matching on WorldCat

88% of manifestations are ‘singletons’ 30% of manifestations are in 12% of the works Average size of multiple matches: 3.1 manifestations/work 43.1 million works in 54 million manifestations 54% of holdings on a FRBR work with >1 manifestation WorldCat manifestations average about 20 holdings

FRBR helps where help is most needed

More FRBR Results

310,000 works have more than 5 manifestations 1.7 million have more than 2 manifestations

Largest: 30,000+ for the Bible 1,537 Shakespeare’s Macbeth 1,026 Dickens’s Christmas Carol

The Top 10 Works by Holdings

Work Holdings Manif’s

1 US Census (various) 403,252 10,164

2 Bible (combined) 271,534 36,738

3 Mother Goose 66,543 1,997

4 Dante, The Divine Comedy 59,034 2,714

5 Homer, The Odyssey 43,871 2,009

6 Homer, The Iliad 42,756 2,388

7 Twain, Huckleberry Finn 39,310 1,093

8 Shakespeare, Hamlet 37,683 1,917

9 Carroll, Alice’s Adventures in Wonderland 37,614 1,865

10 Tolkien, Lord of the Rings 37,461 643

The Top 10 Works Cataloged in 2003

Work Libraries

1 Rowling, Harry Potter and the Order of the Phoenix 2,406

2 Clinton, Living History 36,738

3 Rohmann, My Friend Rabbit 1,997

4 Brown, The Da Vinci Code 2,714

5 Gibaldi, MLA Handbook 2,009

Top 1000 Publication Dates

Top 1000 Languages

Our Beowulf Cluster

24 Nodes• Each with 2x2.6 GHz processors• 4 GBytes memory (96 GBytes total)

One ‘head’ node, 23 ‘compute’ nodes

46x40 GBytes disk (~2 Terabytes total)

Gigabit switch

What we are using it for

All our bibliographic processing• FRBR• Extractions• Searching• Matching

Ganglia load visualization

Starting point

FRBR key generation

25 hours on a 3.00GHz workstation with 2GB of RAM

Generate two key files• sort by key, uniq by key, sort by occurrence• sort by key, post processing on keys, uniq by key, sort

by occurrence

Merge key files

FRBR on the Cluster

44 minutes on the cluster

69 key builders & 23 sort buckets with hyperthreading ON

Generate 23 radix-sorted, post-processed key files

Collapse and sort by occurrence in parallel

Also outputs additional files used by other jobs

Application: Preservation

Identify ‘final copy’ items Do it at the work level

Single-singles• Single manifestations with single holding• Found 18 million in WorldCat

Application: xISBN

A simple Web service

Given an ISBN:• Identify the workset it is in• Return all other ISBNs in that workset

Results should be symmetrical!• Same group retrieved for each ISBN in group

ISBNs sorted by number of library holdings

xISBN Example

http://labs.oclc.org/xisbn/0-19-281664-0 returns:<?xml version="1.0" encoding="UTF-8" ?><idlist>

<isbn>0192816640</isbn><isbn>0820312037</isbn><isbn>0820315370</isbn><isbn>0393015920</isbn><isbn>0393952274</isbn><isbn>0393952835</isbn><isbn>0140430210</isbn><isbn>0192811320</isbn><isbn>0192835947</isbn><isbn>0460872885</isbn><isbn>1853262706</isbn><isbn>0874131219</isbn>

</idlist>

Matching on ISBNs

ISBN additional information beyond Author/Title• Allows relaxation of matching• Introduces possible errors

Offers the possibility of substantial improvement of work matching

Merging Worksets Using ISBN Matches

Pair ISBNs with FRBR keys(Starts with 10 million ISBNs)

Throw out ISBNs in single worksets Throw out ISBNs in > 5 worksets

(We now have 561,000 ISBNs left) Are the titles similar enough? Throw out large groups

Try to be very conservative Authority file always overrides other matching

Matches from ISBN Matching

74,000 author variants ~200,000 title variants

These all create additional cross reference records Automatically folded into FRBR matching Kept separate from NACO file

• Only used in research at this time

Examples of Possible Matches

/mcgraw hill encyclopedia of science & technology /mcgraw hill encyclopedia of science & technology\1\aar aor /mcgraw hill encyclopedia of science & technology\2\apa boo /mcgraw hill encyclopedia of science & technology\3\bor cle /mcgraw hill encyclopedia of science & technology\4\cli cyt …

dickens, charles\1812 1870/tale of two cities dickens, charles\1812 1870/hard times dickens, charles\1812 1870/sketches by boz dickens, charles\1812 1870/martin chuzzlewit dickens, charles\1812 1870/bleak house dickens, charles\1812 1870/little dorrit dickens, charles\1812 1870/oliver twist …

Application: Bookmarklets

Clicking on Princeton

FictionFinder

Indexes fiction from WorldCat Uses FRBR workset algorithm Focused on fiction Searching and browsing by

• Genre• Fictitious Characters• Imaginary Places• Literary Forms

Links to• Google• Open WorldCat

Diane Vizine-Goetz’s project

‘Humphry Clinker’ Search

Work Display

Detail of Language Display

First Few English Manifestations

Manifestation Display

Open WorldCat Link

Additional Matches

Match variant titles:• When the wind blows• When the wind blows: a novel

FictionFinder identified 10,000 of similar variations• novela, novella, roman, …

Created auxiliary authority records Now automatically used when FRBR algorithm is run

Future

Continued development of FictionFinder Extending algorithm to serials? FirstSearch displays Additional matching criteria Local authority files? Integration of auxiliary files for production? Exploring FRBRizing some European catalogs Looking at extending beyond Roman characters

Links

IFLA FRBR - Final Report• http://www.ifla.org/VII/s13/frbr/frbr.htm

Article in DLib• http://www.dlib.org/dlib/september02/hickey/09hickey.ht

ml OCLC Research Activities with FRBR

• http://www.oclc.org/research/projects/frbr/ FictionFinder

• http://fictionfinder.oclc.org/ Top 1000

• http://www.oclc.org/research/top1000/