Extracting Context Using Syntactic Sequences for Research Article Summarization

摘要

Research articles contain information from diverse topics: one of the challenges of creating summaries from research articles is the extraction of information which refers to the main topics. Our summarization approach is to use the contextual information specified in the author's writing style. The contextual information consists of representative terms found by analyzing term frequency in each section and sequences of the terms' grammatical classification called syntactic sequences. We use these document features in our summarization method for the selection of relevant sentences. In addition, we evaluate our system by comparing the sentences extracted by our summarization method with manually selected sentences and also with sentences extracted by other methods.