Contents

• Contents
• List of Figures
• Introduction
  • Scope of the Bilingual Lexicon Extraction Project
  • Why Corpus Linguistics in Bilingual Lexicon Creation?
  • An Overview of this Thesis
• Methods of Bilingual Lexicon Extraction
  • Introduction to Automatical Lexicon Extraction
  • Translating Collocations - Champollion
  • Using Word Alignments - Termight
    • The Termight System
    • Char_align
  • Acquisition of Bilingual Terminology
  • The K-vec Method
  • Automatic Lexicon Evaluation
  • Other Approaches and Techniques
    • BICORD
    • DK-vec
    • Bitext Mapping - SIMR
    • Other Methods
• Lexicon Architecture
  • Norms, Guidelines and Recommendations
    • MULTILEX
    • GENELEX
    • COMLEX
  • MULTRA Lexicon Architecture
• The Scania Project
  • Corpus Characteristics
  • Converting to SGML
  • Sentence Alignment
  • Language Restriction
  • Project Outlook
• Lexicon Extraction from the Scania Corpus
  • Tokenization
  • Cleaning the Database
  • Creating a Basic Dictionary File
  • Searching the Dictionary
  • Looking for Similarities
    • Four Simple Similarity Measures
    • Combining Similarities Scores
    • Length and Word Order
    • Quality of Similarity Scores
  • Evaluation of Dictionary Search Results
  • Cascading the Process
  • Counting Frequencies
    • Word Frequencies
    • Word Pair Frequencies
  • Finding Correlations
  • Low Frequent Words
  • Looking for Compounds
  • Using Part-of-Speech Tags
• Evaluating the Dictionaries
  • Lemmatizing
  • Multiple Translations
    • Why consider Multiple Translations?
    • Filtering with the Dice Coefficient
    • Length Filter
    • Similarity Filter
    • Combined and other Techniques
  • Function Words
• Experimental Results
  • Cleaning the Database
  • Dictionary Search Results
  • Using Similarity Measures
  • Evaluation of Low Frequency Tokens
  • Evaluation of the Dice Coefficient
  • Other Results
• Summary and Conclusion
  • Summary of the Work
  • Concluding Remarks
  • Future Work
• Symbols and Terminology
• Sample for a Cleaning Process
  • Scania Sentence Alignments
  • Cleaned Alignments
  • Modifications and Removed Data
• Theses
• Bibliography