Extractive Document Summarization Using a Supervised Learning Approach

Abstract

In this paper, we present a model for extractive multi-document text summarization using a supervised learning approach. The model uses a convolutional neural networks (CNN) which is capable of learning sentence features on its own for sentence ranking. This approach has been used in order to avoid the overhead of extracting features from sentences manually. Integer linear programming (ILP) approach has been adopted for selecting sentences to generate the summary based on sentence ranks. This ILP model minimizes the redundancy in the generated summary. We have evaluated our proposed approach on the DUC 2007 dataset and its performance is found to be competitive or better in comparison with state-of-the-art systems.