Automatic Text Summarization Using Intenal and Extemal Information

摘要

Sentence ranking is an important step of an extractive text summarization process. Traditionally, sentences are ranked using the information internal to the specified document, that is, they only use the data that is self-contained in the document. External methods use the information that is not necessarily contained in the document, such as manual annotation, labeling, etc. In this paper, we propose a hybrid summarization method that combines the internal methods and the external methods for improving text summarization performance. Comparisons of our proposed summarization system with some state-of-the art summarization systems reveal that our proposed approach is effective.