International Journal of Computer Science Issues

More than a traditional journal...

With the rapid development of the Web, automatic summarization has become more and more important for handling the huge amount of text information in the Web. This paper proposes an automatic summarization method based on compound-word recognition and keyword extraction, termed CASKE. CASKE firstly recognizes the compound-words in a document, labels P.O.S. and revises word segmentation. Then, it extracts keywords, and calculates sentence weights by keyword weights. Finally it selects the proportion of the sentences with large weights to construct summary. The generated summary has good continuity and is readable. Experiment results show that the generated summaries are similar with manual reference summaries, achieving 68.31% Precision and 66.72% Recall in average.