Paper Title: An Assisted Literature Review using Machine Learning Models to Recommend a Relevant Reference Papers List
Author(s) - Ronald Brisebois , Alain Abran , Apollinaire Nadembega* , Philippe N’techobo
This paper proposed an assisted literature review prototype (STELLAR – Semantic Topics Ecosystem Learning-based Literature Assistant Review) based on a semantic metadata ecosystem (SMESE) to discover, rank and recommend the relevant papers for a specific topic. Using text and data mining models, machine learning models and a classification model, all of which learn from researchers’ annotated data and SMESE, STELLAR helps researchers to identify, rank and recommend reference papers for a specific literature review. When ranking a cited document as relevant to the literature review, STELLAR considered many criteria such as venue age, citation category and polarity, researchers’ annotated data, authors’ impact and affiliation institute, and others.
STELLAR algorithms allow to:
- Identify the relevant reference papers for building the literature review from the SMESE, which semantically harvests papers from the web and other sources.
- Obtain the Literature Corpus radius by calculating the distance of each paper to the center of the Literature Corpus defined for a specific topic or area of research.
- Assist the researcher in refining the list of reference papers relevant to a specific literature review.
The performance of STELLAR was evaluated and compared to other approaches using a number of prototype simulations.
Keywords: - Assisted literature review, Machine learning, Semantic topic detection, Text and Data mining.