Interpreting nominal compounds for information retrieval

摘要

A nominal compound, such as computer science, is a sequence of two or more nouns that together form a structure that itself acts like a noun. The interpretation of nominal compounds has been studied in the context of natural language processing and is potentially important for information retrieval. In this paper, we show that simple, knowledge-intensive algorithms can successfully interpret many of the compounds found in technical documents. We also report experiments that indicate that effective identification of concepts represented by nominal compounds may be possible without these knowledge-intensive algorithms.