Mining research abstracts for exploration of research communities

Abstract

Research abstracts are the 'information scents' which attract a novice researcher. To read through the entire research paper and to decide the suitability of the paper to one's research problem is a tough and abstract task. Many times, researchers do not know whether they are citing the relevant (but original?) research articles. It has been only a trial and error approach so far. To enable researchers to correctly target at the relevant and yet quality research literature, mechanisms to organise collections of research papers are essential. Though a considerable effort has been attempted earlier in this context, establishing research communities concentrated on citation based recommendations only. However, the quality and originality of research articles have not been taken into account until now. In this paper, we propose the evolution of research communities by analysing the research abstracts. We utilise Fuzzy Concept Map based approach in detecting the originality of scientific abstracts. By K-means clustering, we establish a research article hyper graph from the qualified abstracts. Later, we evolve the author clusters for every topic cluster and analyse them for redundancy. Further study on relevant bibliometrics helps us to identify a 'nucleus author' for every topic cluster.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.