Analysis of complex network methods for extractive automatic text summarization

Automatic text summarization is an important research area in the domain of information systems. It aims to create a compressed version of documents, which should cover all the significant contents and overall meaning. In extractive text summarization, sentences are scored on various of features. A large number of features network based have been proposed by researchers in the past literatures. This paper reviews all the features that use metrics and concept of complex network for scoring sentences. The experiment results on single feature and combinations of various features we proposed are discussed. Quantitative and qualitative aspects were considered in our assessment performing on the DUC 2002 data sets.