Skip to content

Latest commit

 

History

History
55 lines (46 loc) · 4.63 KB

literature.md

File metadata and controls

55 lines (46 loc) · 4.63 KB
Raw

Litearature Review

Besides search, it can be very useful to just go through the latest papers from these journals and conferences

  • Because relevant papers do not necessarily include the same keywords (a paper about analysing user behaviours may not have the term 'sensemaking')
  • Start from the latest issue/year, and then move onto the year before once completed, and so on.
    • You should at least check the papers from the latest conference
    • Usually one year will be enough. If you want to be more complete, usually the last 3-5 years will be sufficient.
  • Have a quick look at the title and/or abstract to check if the paper is relevant or not
    • many papers can be excluded by the title alone
    • save the paper in Zotero if it seems relevant after reading the abstract
  • The 'Related Work' of these papers will point you to more relevant publications.
  • We will discuss and share these.

For this project, you are likely to find good-quality related work from these journals and conference

What existing work to include and how to discuss

  • Focus on the most relevant: help you to narrow down your scope.
    • For example, if your project is about AI-based recommendations, there is no need to introduce what AI is (the reader would know) and go straight to the latest machine learning models for recommendation.
  • Include academic publications to support your argument:
    • You can include webpages and online articles, but these are not 'peer reviewed', i.e., anyone can potentially say anything online, and as a result these are usually not regarded as reliable evidence.
  • Places to search for academic papers:
  • Things to consider when selecting which papers to cite:
    • IEEE Xplore and ACM Digital Library are usually more rigorous than Google Scholar: all papers published there are checked by one or more researchers, whereas this is not always the case for Google Scholar (such as self-publishing on arXiv).
    • Papers that are more than 5 years old are often outdated (no longer the state of the art), especially in an active research field such as machine learning. Make sure there is no later work that supersedes it if you want to use it.
    • Papers in well-known journals and conferences, most of the ACM and IEEE conferences, are considered strong argument/evidence.
    • You can also get an idea of how good a journal or conference is from its 'impact factor' (has different names) such as the 'h5' metrics used by Google or the 'impact score' by research.com.

How to do citations and references

References are needed for both plan/proposal and dissertation/final report.

  • Two parts: citation in the main text and the references/bibliography at the end of report.
  • Use a tool; don't create references manually.
    • Recommended: Zotero
    • Use Zotero for collecting and inserting citations/references
    • This is a recording of me showing how to use Zotero: https://youtu.be/R_5bjdE-kSw
    • use its browser extension to automatically add relevant paper to your Zotero collection:
      • Zotero connector
      • These can later be exported to bibtex format, if you use Latex.
    • If you are not using Latex yet, use the Microsoft Word plugin to add references to report
  • If you are using Latex,
    • follow this overleaf bibliography tutorial;
    • watch my latex tutorial, which includes references.
    • you can use Zotero to collect the papers from publication website (such as IEEE Xplore and ACM Digital Library) and then generates the BibTeX entry.
    • IEEE Xplore and ACM Digital Library also provide the BibTeX entry on the webpage for each paper: 'cite this' for Xplore, 'export citation' (double quote button) for ACM Digital Library.