Knowledge extraction from research papers to improve their quotations and impacts

The target is to propose a systematic approach to apply tools and techniques of exploitation of the known and published texts in electronic form in order to facilitate the task of creating a good bibliography for research papers, with a particular
focus on leveraging cross-disciplinary citations.

It has to be clearly underlined that the creation of new text mining algorithms is out of the scope of the thesis. The innovative character of this work rather lies in the application of existing approaches to a specific problematic, and its demonstration in a particular area of research.

The methodology and the tools suggested in this thesis make it possible to propose a systematic approach to apply tools and techniques of exploitation of texts effectively for supporting researchers to come up with “good” bibliographies, i.e., with references to literature that is relevant to a proposed paper.

The initial corpus of documents used is the collection of 1500 papers published in the past 10 years in the “CIRP Annals” (Elsevier). The comparison is done based on the analysis of the words extracted from the complete texts and not on simple key words.

The principal utility of this work is to be able to extract a very coherent bibliography relative to the contents of a submitted paper, integrating the transdisciplinarity, and facilitating the innovation by transfer of knowledge from one field to another.

Facilitating the fast diffusion of recent documents can also allow improving the notoriety of a journal by increasing in its impact factor, and for an editor to select more easily the most knowledgeable reviewers.


PhD Student: Pawinee BOONYSOPON
Grenoble University, Defensed on 21st December 2011
Director : Serge TICHKIEWITCH

Currently manuscript is not available.