Monthly Archives: October 2010

Got Big Data Analytic Needs? We Got Big Rapidminer Solutions!

Or should I say, Enterprise Solutions.  I posted yesterday about the amazing amount of answers to questions out there in the forums and blogs for new Rapidminer users.  Often these questions take the form of “Does Rapidminer do “X” or how does the XVal operator work?” Usually from those answers, the user can tinker with their processes and finish up their work.

But what if you’re looking to truly really unleash the power of Rapidminer?

What if your company or special project needs a custom solution that’s open source, stable, platform independent, and has the technical expertise of the Rapidminer developers and power users behind it?

Then you need the Enterprise Solution.  The guys at Rapid-I provide a three tier consulting solution for small to large enterprises, all customized to their needs, and they got a solution for you.

  • Need support setting up automated data mining systems? They do that!
  • Need a custom operator designed? They do that!
  • Want to monitor your company’s brand with out the hassle or servers and software setup? They do that!
  • Need some training? They do that too!

I can go on and on about the depth and breadth of their services and consulting capabilities, but I’ll let their great list of great references ranging from Siemens to Merrill Lynch, and even the Cleveland Indians to speak to that instead.

From Rapidminer’s early incarnation as YALE, I watched this open source software evolve from version 3.4 to 5.0 and I’ve been amazed at its growing functionality and capabilities.  Now, with its new plugin architecture, I can analyze time series data easier than ever before, and text mine Twitter.

It’s been an awesome ride so far and it’s getting better everyday.  Want to come along for the ride?

You Got Rapidminer Questions? We Got Answers

I just wanted to remind my readers, and any new readers, that between the Rapid-I forums, the Rapid-I Blog,  my blog, and my NMT forums, there is an amazing amount of answers to your questions!

First and foremost, the Rapid-I forums and my forums should be your first stop for questions not answered in our growing video library of tutorials.  If that fails, a well thought out (and written) email to yours truly is appreciated. I try to answer every email and forum post (if another reader hasn’t already) I get, but I do a lot of traveling for work and sometimes I forget or don’t answer right away. Please keep pestering me! =)

If you’re just interested in learning about new upcoming functionality in Rapidminer OR just seeing some great ways to extend this software, check out the Rapid-I blog too.  Ingo does and great job showcasing Rapidminer’s existing and future functionality and you should put their RSS feed in your reader.

Genetik Solver appears to be no more – ITS BACK!

Well at least its presence on the Internet. It appears to be gone and I can’t locate the .XLA file anywhere to save my life.  I wrote about this neat little Excel add-in a long time ago in my Genetic Algorithm Excel Addin post and since then its remained amazingly popular.  I sure hope the developer reposts it soon but I’m not holding my breath.

In case you’re interested, there is another slicker Excel add-in for genetic max and min problems from www.xlpert.com.  The only thing is that it costs $35.

FWIW, I actually bought this XLA from the developer because I wanted something to poke around with in Excel without building something in Rapidminer.  Of course you can do the same in Rapidminer, but I’m a tinkerer and a poker and this satisfied my curiosity. :)

UPDATE: The Genetic Algorithm Excel Addin is BACK! The developer moved websites and uploaded it back online again! Thanks Noyan!

A Tweet here, a Tweet there

I’m working with someone to analyze Twitter sentiment tweets.  It’s lookling like a great and interesting project alright and I’ve set up the basic processes already.  I’ll probably build a standard XVal type of model to see what kind of accuracy I get and then weight the positive and negative sentiments with a SVM.  Plus do a lot of data discovery on the classified data.

I fully believe that Twitter is an untapped goldmine for marketers, financial folks, and a slew of all other kinds of people.  The trick to tapping into Twitter and building a viable model first before you start digging around for sentiment.  Why? Because there’s a lot of crap in Twitter too.

They pay us the big bucks for separating the chafe from the wheat…

Rapidminer 5.X Video Tutorial #12 – Using the Generate Attribute Operator to Create Trading Rules

In this video I highlight the often overlooked (I know I did) Generate Attribute operator to create trading rules. This operator is not just applicable to financial data but to any data set where you want to transform and create new attributes “on the fly.”

No sooner did I post this video that I received a notification of a new post at the Rapid-I blog about this operator. They’re going to extend nominal values in the Generate Attribute operator in Rapidminer 5.1.