License: Creative Commons Attribution 3.0 Unported license (CC BY 3.0)
When quoting this document, please refer to the following
DOI: 10.4230/OASIcs.ICCSW.2015.21
URN: urn:nbn:de:0030-drops-54776
Go to the corresponding OASIcs Volume Portal

Gulo, Carlos A. S. J. ; Rúbio, Thiago R. P. M. ; Tabassum, Shazia ; Prado, Simone G. D.

Mining Scientific Articles Powered by Machine Learning Techniques

6.pdf (0.9 MB)


Literature review is one of the most important phases of research. Scientists must identify the gaps and challenges about certain area and the scientific literature, as a result of the accumulation of knowledge, should provide enough information. The problem is where to find the best and most important articles that guarantees to ascertain the state of the art on that specific domain. A feasible literature review consists on locating, appraising, and synthesising the best empirical evidences in the pool of available publications, guided by one or more research questions. Nevertheless, it is not assured that searching interesting articles in electronic databases will retrieve the most relevant content. Indeed, the existent search engines try to recommend articles by only looking for the occurrences of given keywords. In fact, the relevance of a paper should depend on many other factors as adequacy to the theme, specific tools used or even the test strategy, making automatic recommendation of articles a challenging problem. Our approach allows researchers to browse huge article collections and quickly find the appropriate publications of particular interest by using machine learning techniques. The proposed solution automatically classifies and prioritises the relevance of scientific papers. Using previous samples manually classified by domain experts, we apply a Naive Bayes Classifier to get predicted articles from real world journal repositories such as IEEE Xplore or ACM Digital. Results suggest that our model can substantially recommend, classify and rank the most relevant articles of a particular scientific field of interest. In our experiments, we achieved 98.22% of accuracy in recommending articles that are present in an expert classification list, indicating a good prediction of relevance. The recommended papers worth, at least, the reading. We envisage to expand our model in order to accept user’s filters and other inputs to improve predictions.

BibTeX - Entry

  author =	{Carlos A. S. J. Gulo and Thiago R. P. M. R{\'u}bio and Shazia Tabassum and Simone G. D. Prado},
  title =	{{Mining Scientific Articles Powered by Machine Learning Techniques}},
  booktitle =	{2015 Imperial College Computing Student Workshop (ICCSW 2015)},
  pages =	{21--28},
  series =	{OpenAccess Series in Informatics (OASIcs)},
  ISBN =	{978-3-95977-000-2},
  ISSN =	{2190-6807},
  year =	{2015},
  volume =	{49},
  editor =	{Claudia Schulz and Daniel Liew},
  publisher =	{Schloss Dagstuhl--Leibniz-Zentrum fuer Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{},
  URN =		{urn:nbn:de:0030-drops-54776},
  doi =		{10.4230/OASIcs.ICCSW.2015.21},
  annote =	{Keywords: Machine Learning, Text Categorisation, Text Classification, Ranking, Systematic Literature Review}

Keywords: Machine Learning, Text Categorisation, Text Classification, Ranking, Systematic Literature Review
Collection: 2015 Imperial College Computing Student Workshop (ICCSW 2015)
Issue Date: 2015
Date of publication: 23.09.2015

DROPS-Home | Fulltext Search | Imprint | Privacy Published by LZI