Development of Ontology and Thesaurus for Agriculture in China

1
Development of Ontology and Thesaurus for
Agriculture in China

• Li Shijing He Chunpei
• Scientech Documentation and Information Centre
• Chinese Academy of Agricultural Sciences
• Beijing 100081, China

2
1. Chinese Language Agricultural Thesaurus

• passed the ministry level identification in1992
• printed in hardcopies in July 1994
• won the State Third Prize of Scientech Progress
  awarded by State Science and Technology
  Commission in 1995

3
1.1 Main achievements in the development of
Agricultural Thesaurus

• Selection of terms
• a tool in document indexing and information
  retrieval ------ a component of the State
  Thesaurus Database
• We have expanded the selection of terms to
  include agricultural information retrieval
  systems, relevant thesauri, relevant authority
  dictionaries, textbooks, monographs, periodicals,
  related classification schemes and terms
  recommended by agricultural experts and users.

4

• Capacity Agricultural Thesaurus (AT) includes
  240 domains in 45 disciplines and contains 63,000
  entries, with 51,000 descriptors and 12,000
  non-descriptors, the largest agricultural
  thesaurus in the world. AT just meets the demand
  of rapid development of modern agricultural
  science in the world.

5
1.2 Enhanced compatibility

• (1) AT is compatible with both foreign and
  domestic thesauri
•  
• (2) AT is compatible with the Chinese
  Classification Scheme
•  
• (3) AT is compatible with the State descriptor
  database.

6
1.3 Key technologies and innovation points

• Thesaurus mode and display design AT has adopted
  integrated mode of classification and subject and
  full display of interterm relation

7

• (2) Construction of descriptor reference system.
  It is the key in thesaurus development. AT has
  the following innovations
• ? Only establish three kinds of relations used
  for, related term and narrow term
• ? Use reference relation to narrow down big word
  family relation, hence, enriched the
  compatibility and inter transformation capacity
  of AT, and improved thesaurus quality and
  searching efficiency.

8

• (3) Thesaurus compatibility improvement
  technology
• Innovation point unified descriptor and
  classification names, thus made AT to be
  compatible to different kinds of search languages.

9

• (4) Compilation method
• Innovation points
• Change the traditional method to
  integrated domain as the unit, thus increasing
  the efficiency by 3 times.

10

• (5) Use computer to realize automation in
  database establishment, table compilation,
  management and printing,
• With the following innovations

11
Innovations

• ? Used the third model to design the thesaurus
  database so as to solve the difficult issues of
  interterm relation.

12
Innovations

• ? Establish special Chinese character database
  and realize comprehensive sorting of Grade I and
  Grade II character database.

13
Innovations

• ? Establish maintenance system of compilation and
  Chinese character database.

14
Innovations

• ? Realize linkage of multiple systems. We can
  generate MINISIS system database and also machine
  readable thesaurus in laser-jet printer.

15
Innovations

• ? Establish thesaurus output system, which is
  selective and general use. It can generate
  thesaurus of multiple structures.

16
1.4 Users Comments

• (1) AT has wide domain coverage with broad
  collection of terms, standardization and high use
  rate.
• (2) AT has a rational general structure, which
  has expanded view when query.
• (3)  AT has comprehensive interterm relations and
  adopted full display mode, which is convenient
  for query, thus improved success rate in first
  reading and saved time in indexing.
• (4)  ATs offsetting is compact with page headers
  indicator, which has provided convenience for
  query.

17
1.5 Social and Economic Benefits

• AT has been included in the State Descriptor
  Database, which will bring into full play the
  role of AT in the modernization of agricultural
  information work. Under the support of the
  computer compilation system, we have produced the
  machine-readable product and hardcopy of AT. In
  this work the work efficiency has been improved
  by 300 and saved 75 of the production cost.

18
Major social and economic benefits

• (1) AT can provide unified, standard tools for
  the users in information resources sharing.

19
Major social and economic benefits

• (2) Due to the standardized terms in AT, it
  provides good software environment and data
  preparation for the standardization work, such as
  Agricultural Terms Database.
•  

20
Major social and economic benefits

• (3) Due to the fact that AT is a multilingual
  thesaurus, it can be used in network for
  international information resources sharing.

21
Major social and economic benefits

• (4) AT can be used in any machine translation
  systems, thus promote automatic indexing of
  agricultural documents and machine translation.
•  

22
Major social and economic benefits

• (5) AT can provide good conditions for providing
  common vocabulary for agricultural institutions,
  research institutions and universities. AT can
  also be used in office automation of these
  institutions and universities.

23
Major social and economic benefits

• (6) On the basis of the classification tables of
  AT, we can develop and research on classification
  schemes of different agricultural domains, thus
  making even greater contribution for agricultural
  information searching language and theory.

24
Major social and economic benefits

• (7) We can develop agricultural and biological
  vocabulary compendium on the basis of biological
  name classification database and agricultural
  descriptor database, so as to standardize
  agricultural terms and biological names.

25
Major social and economic benefits

• (8) AT has been used by 30 academies of
  agricultural sciences and more than 40
  agricultural universities in subject indexing and
  information query when they create agricultural
  information databases. AT has made great
  contributions in the standardization of indexing
  and query in modernized networking age.

26
2. Application of Agricultural Thesaurus

• ----- Study on computer translation system
•  

27

• SDIC/CAAS has developed a computer translation
  system on the basis of AT.
• The system has a dictionary database with a total
  number of over 300,000 words.
• The system has been released to public in 2000
  and is now broadly used by clients in various
  economic circles.

28

• The machine translation system includes
  following domains
• Crop industry, animal science, biology,
  biotechnology, plant pathology, animal pathology,
  soil science, soil conservation, food technology
  and biological classification.
• It also contains words in chemical engineering,
  computer science, genetics and 72 domains related
  to agriculture.

29
2.1 Content of the study

• The main study contents include
• ? Collect agricultural English common vocabulary
  and create an English-Chinese agricultural
  electronic dictionary

30

• The main study contents include
• ? Study and compile agricultural language
  material databases for both translating English
  into Chinese and vice versa.

31

• The main study contents include
• ? Select machine translation software packages
  through comparative studies.

32
2.2 Significance of the study

• ? It is an important means in making use of
  international agricultural information resources.
  It has promoted the Chinese agricultural
  scientech information to go to the world and for
  the formation of a Chinese agricultural
  information industry.

33
2.2 Significance of the study

• ? It is of great importance to improve the
  efficiency of English-Chinese and Chinese-English
  translation so as to meet the demand of
  translation service.

34
2.2 Significance of the study

• ? The study has not only promoted the unfolding
  of agricultural term study in the world and also
  opened up a new area in computer application.

35
3. Application of Agricultural Thesaurus

• ----- Translation of AGROVOC multilingual
  thesaurus
•  

36

• FAO published the first edition of AGROVOC in
  1982, and then published the second, third and
  fourth editions in 1988, 1995 and 1999,
  respectively.
• FAO had issued the network edition in 2000.
• The fourth edition had four languages, namely
  English, French, Spanish and Arabic.

37

• AGROVOC has a total number of 16,607 descriptors
  and 10,758 non-descriptors, which are used for
  indexing information materials produced within
  the international cooperative information system
  AGRIS and CARIS, and for data retrieval from
  those systems.

38

• We first used the AT database to screen the
  descriptors in AGROVOC, and found more than 50
  of the descriptors were already in AT.
• Therefore, the translation load has been reduced
  half when we used the Agricultural Thesaurus.

39

• There are about 20 agricultural domains in
  AGROVOC. All the descriptors and non-descriptors
  need to be divided according to their specialties
  and distributed to the translators before
  translation.

40

• It is impossible to divide it based on English
  and their term code. We found all the first terms
  in every word trees based on AGROVOC word
  relation and created a new database, which
  included every word tree.

41

• The translator can translate the term starting
  from the first level to the narrower terms till
  the last level.
• In every hierarchical level, we also put out
  their related words, but translators need not
  translate these words, because all these words
  will appear in the hierarchical levels.

42

• As some terms belong to different word trees,
  therefore, there are some overlapping terms in
  our database.
• The total number of overlapping terms is about
  2000, which may be allocated to one or several
  translators

43

• When the same terms in different word trees were
  translated, we can find very small differences
  among their translations and can select a term,
  which is more appropriate in the translation.

44
4. Outlook on information retrieval in networks

• SDIC/CAAS is a government institution, which is a
  national research organization in the field of
  agricultural information research in China.

45

• SDIC/CAAS is looking for both research partners
  and funds in agricultural ontology development.
  We are willing to cooperate with any
  international or national organization.

46

• China is willing to cooperate with any
  international or national organization so as to
  facilitate more capable results in information
  network search.

47
(No Transcript)