Research Topic Detection Using a TV-tree Based System

Abstract

There is a tremendous growth in scientific research, thereby resulting in a huge number of research articles being published. Most of these articles are available online and are accessed using various search engines and online repositories. Selecting the appropriate search keywords facilitates an efficient search from this pool of articles. Naive researchers find it difficult to choose such keywords and are unable to identify articles relevant to their topics of interest. This paper proposes a tree-based approach to detect research topics from a set of research articles. This enables naive researchers to choose appropriate keywords and to identify articles relevant to their topics of interest and in turn pick the trendiest topic for their research. The topic detection system proposed herein is experimented with articles in computational geometry, a branch of computer science that deals with study of algorithms. Experimental results reveal that the top influential research topics in the area of computational geometry are detected efficiently by means of the proposed topic detection system.