For Python coders, most mainstream machine learning algorithms are implemented in scikit (save for genetic algorithms, which have their own popular implementations). What's a good Java alternative that doesn't conflict with a developer's ability to copyright commercial software that uses machine learning for some of its problem-solving? Although JSAT supports many of the same algorithms as scikit, it's unsuitable due to a GPL licence.

  • The GPL doesn't prevent you from copyrighting software that you write with a GPLed library. All that is prohibits is restrictively relicensing the software, if you distribute it or a derivative work of it. If you merely write your program to interface with JSAT and don't distribute JSAT with your program, you should be clear to license your program however you want. – Kodiologist Oct 20 '17 at 19:35
  • 1
    @Kodiologist For my purposes that's a difference without a distinction. – J.G. Oct 20 '17 at 19:38
  • The Apache licence is very permissive and the Apache software foundation has a number of Machine learning projects Apache Spark is very popular as well as it's machine learning library MLlib. I'm not writing this as an answer because my knowledge in this area isn't authoritative enough. – Quaternion Nov 3 '17 at 1:42
  • Siri, Alexa, et. al. could be under the GPL but still unusable. Why? Because the missing pieces aren't and never would be distributed. Take away for you: run it in the cloud. No distribution, so no compliance problems. – Thufir May 25 '18 at 2:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.