Advancing Your Topic

• Finding sources
  • We will continue to look for the latest in cluster analysis.
    • Start by a google scholar search for k means cluster analysis
    • This yields three very interesting results at the beginning
      • An efficient k-means clustering algorithm: Analysis and implementation.
        • This is interesting because
          • It has over 3000 citations.
          • It is over 30 pages, and provides a very strong background in the subject.
          • It has a nice empirical section
          • It has good results.
          • In my opinion it is very readable.
        • The major drawback is it was published in 2002, somewhat dated.
        • But this would definitely go in my collection of papers
        • Technical for soon:
          • I would save the file as kanungo2002.pdf
          • In the folder I would add the Bibtex information

            @article{kanungo2002,
              title={An efficient k-means clustering algorithm: Analysis and implementation},
              author={Kanungo, Tapas and Mount, David M and Netanyahu, Nathan S and Piatko, Christine D and Silverman, Ruth and Wu, Angela Y},
              journal={IEEE transactions on pattern analysis and machine intelligence},
              volume={24},
              number={7},
              pages={881--892},
              year={2002},
              publisher={IEEE},
              pdf = {kanungo2002.pdf},
              url = {http://ieeexplore.ieee.org/abstract/document/1017616/},
              annote = {This is an old document, but it has a very nice write up and has been cited many times.  The empirical section has good results and it would be very interesting to attempt to reproduce.}
            }
            

          • The pdf and url entries are not used in the bibliography but are for my record keeping.
          • The annote field is for the annotated bibliography package I am using.
      • Data clustering: 50 years beyond K-means
        • This is interesting because
          • It is cited over 3000 times as well
          • It has a nice overview of different clustering algorithms
          • It has two pages of bibliography.
          • It is reasonably new (2010)
          • It has some interesting results, including cluster analysis done on a tattoo.
          • This looks to me to be a precursor to a book chapter or perhaps even a book.
        • From this paper, I did a google search for the author.
          • His home page linked here
          • He has MANY publications
          • It looks like he is interested in face and fingerprint recognition.
          • Not quite what I wanted, but worth a look.
        • I also looked at the citation link on this page.
          • I like the looks of A comparative study of efficient initialization methods for the k-means clustering algorithm.
            • This is 2013
            • And gives insight into one part of the problem.
            • There are others here, including a 2011 book.
          • I searched for articles from 2015 on.
          • This will give me an overview of what is new.
          • Save this for later.
      • Algorithm AS 136: A k-means clustering algorithm.
        • This is very old (1979)
        • It contains reference code, in FORTRAN
        • Probably a standard reference.
        • I would save this and read it due to
          • The age of the article
          • The number of citations.
  • CiteSeerX
    • A search engine.
    • I liked the previous version much better.
    • But you may have some luck here.
• Remember,
  • Search for other works by authors of useful articles.
  • Look at all the authors, not just the first listed.
  • Look at home pages, graduate students and so on.