2

I'm searching for books on the basic k-means and divisive clustering algorithms. I'm interested in the pros and cons of both. It's a part of my bachelors thesis, I have implemented both and need books to create my used literature list for the theoretical part. Also, is there a book on "the curse of dimensionality"? Thanks!

MustSeeMelons
  • 149
  • 1
  • 5
  • 1
    What books/sources did you actually use to find out about the techniques and implement them? This is what the literature should be about, I think. Padding it with stuff you didn't read makes no sense. – Gala Apr 29 '13 at 12:14
  • I used one book in my native tongue. I have checked: Data clustering: theory, algorithms, and applications. Data mining: concepts, models, methods and algorithms and Cluster Analysis, 5th edition. I don't need no padding, just a few books in which the algorithms are well described, with their pros and cons. For example: 1) the centroids in k-means tend to move closer to where there are more samples. 2) Due to the "cure of dimensionality", it's only possible to compare results with in the same data set. I just need ground to stand on, that I read that in a book, not thought of it myself ;) – MustSeeMelons Apr 29 '13 at 12:31
  • 1
    There is nothing wrong with crediting a book in your native tongue, **if that is what you used**... – Has QUIT--Anony-Mousse Apr 29 '13 at 20:12
  • This was genius. Very useful. – MustSeeMelons Apr 30 '13 at 06:27

2 Answers2

2

Almost any book on machine learning will explain you the so called "curse of dimensionality" problem. For example, the freely available and excellent Information Theory, Inference, and Learning Algorithms. It also has a section on clustering algorithms, though probably not in all depth you may wish for.

jpmuc
  • 12,986
  • 1
  • 34
  • 64
-2

Here is a book that's got a couple of them + very simple examples.

HTH

D

dfhgfh
  • 401
  • 1
  • 4
  • 9