Key points are not available for this paper at this time.
This paper describes a new method for computing lexical chains. These are sequences of semantically related words that reflect a text's cohesive structure. In contrast to previous methods, we are able to select chains based on their cohesive strength. This is achieved by analyzing the connectivity in graphs representing the lexical chains. We show that the generated chains significantly improve performance of automatic text summarization and keyphrase indexing.
Olena Medelyan (Mon,) studied this question.