Variant personal names for the same Japanese historical individuals exist, and when handling historical data it is desirable to control these. Furthermore, by grasping the position the family an individual belongs to within a genealogy or organization, it is possible to estimate the individual’s social position and the power he might command. At present there is no database providing such information in Japan, and there is a need to construct the authority information for personal names structured in a standardized data descriptive language. On this basis, the present study describes a project to construct authority information for the former Japanese noble families, which played a central role in the modernization of Japan, and for persons related to them, using topic map.
1. TMRA 2009
Construction of Authority Information for
Personal Names Focused on the Former
Japanese Nobility using a Topic Map
p y g p p
2009/11/12, Leipzig, Germany
, p g, y
Norio Togiya
(togiya.norio@iii.u-tokyo.ac.jp)
University of Tokyo
Motomu Naito (motom@green.ocn.ne.jp)
Knowledge Synergy Inc.
2. Table of Contents
1. Introduction
2. Target of investigation
g g
3. Constructing Authority Information
4. Demo (authority topic map)
5. Issues and Discussion
5.1 Person s
5 1 Person’s name problem
5.2 Diversity of Information Items
5.3
5 3 Problems of Centralized Topic Map
6. Future work:
Toward Distributed and Linked Topic Maps
7. Conclusion
3. 1. Introduction
Background
・ There are many variant p
y personal names for the same Japanese
p
historical individual
・ When handling historical data it is desirable to control them
・ But there i no database providing such information in Japan at
h is d b idi hi f i i
the present
・ There is a need to construct the authority information for personal
names structured in a standardized data description language
Purpose
・ To investigate and analyze persons who played significant social
and cultural role
・ To construct the authority information of them to support
historical and cultural study
4. 2. Target information
・ In the first stage, we are constructing a topic map of the authority
information relatively small scale and limited area
・ We are focusing on the former Japanese nobility
・ Japanese aristocracy is existing from after Meiji Restoration in
1869 until after the end of WWⅡ i 1947
til ft th d f in
・ They played significant social and cultural roles in the
pre WWⅡ
pre-WWⅡ period
・ They often changed their name and had many alias names
・ Meanwhile different persons often had the same name
5. 3. Constructing Authority Information
We constructed our first topic map for the Authority
Information according to the following process
I f i di h f ll i
- Categorizing authority information
-O l
Ontology making
ki
- Topic map making
- A li i making
Application ki
6. 3.1 Categorizing authority information
・ We collected and analyzed information items
・ We categorized those items and mapped them to information items
of Topic Maps
・ The following table shows the categories and TM correspondence
table: Categories of personal name source data (1/3)
Categories of personal name authority information Correspondence in Topic Maps
Name Kanji (family name/personal name) Topic name
(multiple responses Reading (family name/personal name) Variant and/or Internal occurrence
possible) Romanization (family name/personal name) Variant and/or Internal occurrence
Type of names (alternatives or childhood Variant and/or Internal occurrence
names) (multiple responses possible)
Nationality (multiple responses possible) Linked by association to other topics
Gender (multiple responses possible) Linked by association to other topics
Rank (multiple responses possible) Linked by association to other topics
Profession (multiple response possible) Linked by association to other topics
Person ID Subject ID
7. table: Categories of personal name source data (2/3)
Categories of personal name authority information Correspondence in Topic Maps
Related URL/URI Person URI External occurrence
Related URL (multiple response possible) External occurrence
Dates of birth and DOB (Western calendar only) External occurrence
death (multiple responses possible)
DOD (Western calendar only) External occurrence
(multiple responses possible)
Brief biography Japanese biography Internal occurrence
English biography Internal occurrence
Place of birth (multiple responses possible)
Pl f bi th ( lti l ibl ) Linked by
Li k d b association t other t i
i ti to th topics
Place of residence (multiple responses possible) Linked by association to other topics
8. table: Categories of personal name source data (3/3)
Categories of personal name authority information Correspondence in Topic Maps
Administrative data Date of input (multiple responses possible) Internal occurrence
Last update Internal occurrence
Type Internal occurrence
Language code (multiple responses possible) Internal occurrence
Character code Internal occurrence
Source confirmation Internal occurrence
Input by (multiple responses possible) Internal occurrence
Relationship (multiple Teacher, student, acquaintance, father, mother, Association
responses possible elder brother, elder sister, younger brother,
younger sister, husband, wife, child
9. 3.2 Ontology making
We made ontology according to the categorized items (subjects)
and relationships between them
Ontology diagram of the topic map
- Squares represent Topic types
- Lines represent Association types
10. 3.3 Topic map making
- The topic map was generated using DB2TM which is
included in Ontopia
p
- Ontology definition file and XML configuration file are needed
for DB2TM
- Ontology definition file defines the following:
- Topic types
- Name types
- Association types
yp
- Association role types
- Occurrence types
- XML configuration file defines the mapping rule from EXCEL
(CSV f format) i
) into the ontology d fi i i
h l definition
11. 3.4 Application making
We developed the application using Ontopia Navigator Framework
The f
Th feature of the web application
f h b li i
- Displaying instance list of each
J2EE Web Server
topic type e.g. Tomcat
- Displaying instance detail http
(names, occurrences and assciations) JSP Page
- N i i topic map
Navigating i topic
- Character string search map
Taglibs
- Tolog query interface
- Graphical representation <HTML>
pages
Query engine
server client
(Source: Ontopia, “The Ontopia Navigator Framework Developer’s Guide” )
12. 4. Demo
The b
Th web application for personal authority topic map
li i f l h i i
Screen shots of the application
13. 5. Issues and discussion
5.1 Person’s name problem
- Many names for one person
y p
- The same name for many persons
- Three notations for each name
Kanji name
Reading (Katakana or Hiragana name)
g( g )
Roman name
- How to describe them as topic name
p
- Content model is showed as follows:
name = element name { typicalName, aliasName* }
typicalName = element typicalName { kanjiName, katakanaName, romanName }
aliasName = element aliasName { kanjiName, katakanaName, romanName }
j
14. 5. Issues and discussion
5.2 Diversity of Information Items
5 2 Di it f I f ti It
(1) Two kind of information items
・ Fundamental information items
d li f i i
They are good candidate for PSI and PSD
ex: typical name alias nationality gender, orders,
name, alias, nationality, gender orders
date of birth and death, born and lived place, etc.
・ Specific information items
p
They change according to individual domain and view
ex: biographical outline, achievement, personal connection,
position, expertise, etc.
ii i
(2) Items not depend on person
ex: place country organization, occupation, etc.
place, country, organization occupation etc
・ We cannot make exhaustive list for them if we pick up them
by occurrence basis. But if we make those list once, we can
y
share them among many application
15. 5. Issues and discussion
5.3 Problems of Centralized Topic Map
・ Authority information consists of diverse items and many
independent items
・ It is very difficult and troublesome to integrate those items into
one centralized topic map
・ Such topic map become complicated, hard to understand and
difficult to maintain
・ Moreover there are different relations depending on domains
and ranges and they change according to the point of views
・ It is desirable that we can filter out specific relation and link
from others flexibly
16. 6. Future work:
Toward Distributed and Linked Topic Maps
Instead of centralized topic map, distributed and linked topic maps
are preferable
・ Those topic maps are specialized and relatively simple and small
・CCurrently a large amount of person’s information is inherited by
tl l t f ’ i f ti i i h it d b
many libraries, museums, research institutes, etc. separately.
・ We think it is natural those organization continue to manage them
・ We are making topic maps about information owned by them
- Author information owned by National Diet Library:
800,000 records
- Historical person information owned by National Institute of
Japanese Literat re: 50,000 records
Literature: 50 000
・ Next we plan to create topic maps for places, countries,
organizations, occupations, etc individually
・ Then we will make effort to link them
17. Toward Distributed and Linked Topic Maps
We
W are planning to use the mechanism of TMRAP, Subj3ct,
l i h h i f TMRAP S bj3
Ontopedia to realize the Distributed and Linked Topic Maps
18. 7. Conclusion
・ As the first stage, we created the topic map for personal name
authority information focused on the former Japanese nobilities
y p
・ It made clear the genealogies, the network of the marriage
and other interrelationships between them
・ We believe our authority information is very useful for
researchers to study persons and their network related social,
social
cultural and historical affair
・ There are strong needs to personal authority from various domain
・ The data structure, Topic Maps, and the system structure we
propose have generality scalability and flexibility
generality,
・ Thus, those are adaptable for various fields in the future