Analysis of the New Zealand Emglish and Mori OnLine Translator

1
(No Transcript)
2
Analysis of the New Zealand English and Mäori
On-Line Translator

• Mark Laws, R Kilgour, M Watts
• Knowledge Engineering Laboratory
• Department of Information Science
• University Of Otago
• Dunedin
• New Zealand

3
Introduction

• The Translator provides on-line web users with a
  bilingual interface
• The Translator is a product of the Otago Speech
  Corpus Project
• The Translator is an integrated relational
  database system
• On-line analysis has identified the Translators
  future direction

4
Web Development

• The web is an effective environment to test the
  Translator
• All tools used to develop the web based system
  are open source
• Linux Redhat 6.0
• PostgresSQL
• Apache HTTP Server
• HTML PHP Scripting
• Documentation, Datasets, Static Active Links

5
(No Transcript)
6
(No Transcript)
7
A Bilingual System

• Both English and Mäori text can be displayed
• Instructions
• Messages
• Dialogues
• Comments
• The web interfaces are bilingual
• Both languages have equal status

8
Web Page Analysis

• Two main log files are generated from the
• Apache server
• PHP-Postgres scripts
• Five important areas identified
• Web Domains (eg. com, net, nz)
• English translations
• Mäori translations
• Bilingual Interface usage
• Speech files accessed (.WAV )

9
Translated Words

• Submissions (13k E, 8.7k M)
• Translations (68 E, 43 M)
• Mis-hits (high - nil data, spelling)
• User Errors (high - spelling, orthography,
  interface)
• Unsuccessful translations high
• Lexicon size (more words, phrases)
• User interface perception (incorrect method)
• User mis-spelling (esp Mäori words)

10
Mäori Speech

• Provides reinforcement to the Mäori translations
• Givens the correct spelling and pronunciation
• 50 of translated words were accessed played
• This suggests that speech translation is important

11
Summary

• Proves that a specialised web based translator
  offers more user interaction
• On-going analysis provides up-to-date
  improvements to the system
• Justifies future directions for adding speech
  interfaces and spelling tools
• Next version implemented using Java

12
Additional

• Translator http//kel.otago.ac.nz/translator/in
  dex.html
• Datasets http//kel.otago.ac.nz/maaka/dataset
  s/
• This paper http//kel.otago.ac.nz/maaka/datasets/t
  ranslator.pdf
• On-line paper http//kel.otago.ac.nz/maaka/dataset
  s/translator.html
• This presentation http//kel.otago.ac.nz/maaka/dat
  asets/presentation.ppt

13
Acknowledgements

• Information Science and Mäori Studies Colleagues
• New Zealand Foundation for Research Science and
  Technology - CBIIS Grants and Fellowships
• JCIS 2000 Committee

14
(No Transcript)