资 源 简 介
The core of this project is to develop a prototype being able to determine the most possible relationship between the constituents of English compounds. This functionality could be used for machine translation or spell checking. As a fairly new approach this project takes advantage of heavy querying of search engines with pre-defined phrases, each representing a possible relationship. In the project"s scope search engines act as an interface to the knowledge of the world in modern day internet. Our concept is based on the assumption that if a search engine finds a high number of results for a phrase, it is commonly used and widespread in natural language and therefore "correct" in the interpretation of a human speaker.
The prototype developed to prove this idea and problems we"ve faced during the project,
from decompounding, resolving relationships between compounds, to the art of defining
reasonable search queries, are discussed in this paper.
The prototype"s accu
