Burst Analysis of Text Document for Automatic Concept Map Creation

Abstract

In this paper, we propose a new method to extract relationships between words based on the burst analysis for creating a concept map from a document. Concept maps are graphical representation showing the relationships among concepts. An automatically generated concept map shows the whole picture of a certain domain or a document and helps people to understand it. A traditional approach to capture the association relationship between concepts uses co-occurrence of words. In this approach, the highly frequent words usually have strong relation with other words. However, these relations do not necessarily describe the content precisely. Instead of counting co-occurrence of words, the proposed method analyses burst interval of a word for detecting a topic word in a particular period and captures the relation between burst intervals. The case study shows that the proposed method outperforms the co-occurrence method in ranking meaningful relation-ships highly. © 2014 Springer International Publishing Switzerland.