Semi-Extractive Text Summarization Approach to Extract Research Objective of Academic Literature

Abstract

The task of literature review is an important for all research areas. The volume of academic literature published in the scientific journals, books and conference proceedings is increasing rapidly. It has become very difficult for researchers to analyze huge amount of published research papers and draw conclusions. Therefore, an automatic system able to extract the research objectives that recap the meaning and the definition of research from the article published is needed. The main challenge is to deal with the highly multiplicity of authors’ writing style used in forming the objective sentences. A semi-extractive approach for precisely extracting research objective from the research abstract is proposed. The extraction process is performed in two tasks, search for pre-collected Key Expressions and verify the findings using pre-built General Key Phrases. An extensive evaluation has been conducted to find out how well it is performing using special dataset that includes diversified writing styles. The results showed that the method is a promising and able to extract objective sentence accurately.