Phrase processing for detecting collocations with KoKS

Norman Kummer and Joachim Wagner (2002): Phrase processing for detecting collocations with KoKS, In online Proc. of Colloc02 Workshop on Computational Approaches to Collocations, http://www.ai.univie.ac.at/colloc02/, Vienna, Austria

• Handout [624 KB]
• Presentation Adobe PDF [2176 KB]
• Paper [880 KB]

Abstract

These paper presents an approach to extract semantically opaque collocations from parallel corpora. Possible translation correspondences on phrase level are automatically extracted and ranked by a measure of literalness. We assume that non-literal translations indicate non-compositional semantics of a phrase. The paper discusses corpus preprocessing, alignment and collocativity measure, presents results and discuesses some problems.

Table of Contents

Some headings shortend or improved

1. Introduction
2. Definition of Collocations
3. KoKS System
   1. Corpora and Dictionaries
   2. Normalisation
   3. POS Tagging and Lemmatisation
   4. Sentence Alignment
   5. Phrase Alignment
   6. Detection of Collocations
   7. Database
   8. Demo-Application
4. Results
   1. Phrase Alignement
   2. Collocativity Measure
5. Proposed Applications
   1. Collocation Dictionary
   2. Providing Examples
   3. Translation Memory
6. Related Work
7. Outlook and Open Problems