-1

I would like to install ELKI on windows and use clustering methods from python source code. The official documentation on the installation and use of their methods is not clear, which makes it difficult to handle their tool.

pedro colombino
  • 27
  • 6

1 Answers1

0

It is a Java program. It does not need to be "installed".

If you have Java, you can just launch the jar file.

But the target audience are scientists and developers, these know how to run a Java program, and will want to check out the source code from Git anyway.

Erich Schubert
  • 8,575
  • 2
  • 26
  • 42
  • thank you for your response. It is not documented enough to use it in source code (I'm not talking about GUI where just running a jar file is enough). – pedro colombino Feb 15 '20 at 09:59
  • There is extensive JavaDoc documentation; and backwards search in your IDE will give you many usage examples. But of course the code and APIs are fairly advanced, and try to give high type safety (via generics and such) rather than making it as easy to use as guessing. But it does not make much sense to write a lot of "beginner documentation" when its not a version 1.0 yet and APIs still change a *lot*. Much of the documentation is already outdated (and usage has become easier, not harder, since). But **contribute documentation** that you consider missing! – Erich Schubert Feb 15 '20 at 20:42
  • I would do it with pleasure after a good mastery of Elki that I find interesting. And why not set up an API for python because a lot of data scientists use python for their projects. – pedro colombino Feb 17 '20 at 10:29
  • The target audience are people that want to modify algorithms to improve them, for these a Python wrapper does not offer much benefits. You can use Python to prepare your data, then launch ELKI to run the algorithm, and read the result back into Python; there is not much to be gained from a closer integration. – Erich Schubert Feb 19 '20 at 20:55