ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Construction and exploration of networks of scientific papers

Niko Colnerič (2013) Construction and exploration of networks of scientific papers. EngD thesis.

Download (1480Kb)


    We present an innovative approach to reviewing related scientific articles. Rather than showing articles in a list, we show related articles in a network. Edge width represents semantic similarity of connected articles. We define similarity between two articles as cosine similarity between vectors, which represents this two articles. Vectors are obtained by TF-IDF transformation of article abstracts. Before transformation abstracts are first lemmatized. Information about articles are gathered from online repositories PubMed and CiteULike. We use caching to speed up the access. We also present a method of removing less important edges and techniques used to visualize this networks with D3 library. Our application is publicly accessible on website http://butler.fri.uni-lj.si/articles.

    Item Type: Thesis (EngD thesis)
    Keywords: text similarity, visualization, network of scientific articles
    Number of Pages: 44
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    izr. prof. dr. Janez Demšar257Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=51012&select=(ID=10092628)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 2124
    Date Deposited: 03 Sep 2013 16:43
    Last Modified: 20 Feb 2015 11:00
    URI: http://eprints.fri.uni-lj.si/id/eprint/2124

    Actions (login required)

    View Item