Extractive Text Summarization

The aim of the text summarization is to extract the most important pieces of information from the text. Another substantial aim is to summarize the text with consistent, representative information. Summaries by hand have been quite successful. However, creation of summaries by hands is not possible because of huge amounts of text data. Thus, it is necessary that the text data is summarized automatically.  Automatic text summarization enables users to gain brief, consistent and representative summaries. Moreover, it provides that the reading time for the users decreases and the selection process of the documents is easier for users.

In this project, I have studied on the effect of word embedding and phrase embedding on extractive multi-document text summarization.  Two types of sentence selection methods which are Submodular Function Optimization and LexRank are implemented.  All combinations with summary framework and sentence representations are evaluated.