1
$\begingroup$

I love Weka for its visual data exploration features (it's great at quickly giving you a mile-high aerial view of the distributions of all your variables).

I have a 0.5gb dataset, and Weka is unable to load it (on a 64-bit Windows PC with 16gb of RAM).

For the Weka fans out there, what's the closest 'visual data exploration' alternative for very large datasets?

(FWIW - I've done all the various things that are recommended when Weka fails at loading a very large dataset, e.g., increasing the Java heap size, etc.)


Update: I tried RapidMiner Basic (it's open source and free), and that worked.

$\endgroup$
1
$\begingroup$

I tried Azure Machine Learning (by Microsoft) and it is very nice. It can easily handle large data sets and has similar to weka pipeline for ML tasks and visualisation. They offer it free for 30 days so you can try it. You can google it or follow this link Azure ML

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks for the idea. I don't want to use trialware, but that may work for others out there. I tried RapidMiner Basic (it's open source and free), and that worked. $\endgroup$ – thanks_in_advance Apr 16 '16 at 0:10
1
$\begingroup$

First I wanted to propose Orange Data Mining Suite.

enter image description here

But then, I know almost for certain that there isn't a Windows version of 64bit NumPy available, which Orange uses, and thus your data probably won't fit in with the 32bit Windows release they provide out of the box. It may be installable if you can find your way around Python well enough, but it may just also prove being easier to switch onto some more productive platform (i.e. GNU/Linux and Macintosh OS X are supported).

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks, appreciate the ideas. I did try Orange already, and you're right, that didn't work on my Win PC either. Here's what seems to be working: I tried RapidMiner Basic (it's open source and free), and that worked. $\endgroup$ – thanks_in_advance Apr 16 '16 at 0:08

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.