Go to the profile of Martin Schmitz, PhD
Martin Schmitz, PhD
Free Thinker and Data Scientist writing about data and business life.

7 Things You Need To Do After Installing RapidMiner 7.2

RapidMiner 7.2 is now round about 2 month old. Time to have a look what you should do after installing RapidMiner.

Try out Radoop it’s free now!

With the latest release RapidMiner Radoop got a free version. It is not only free, but also powerful. You can builtyour big data analytics in minutes using a code-optional GUI. The free version supports both MapReduce based data wrangling but also Spark machine learning.
You don’t have a cluster at hand? You can use a Cloudera or Hortenworks Sandbox to try it. On RapidMiner’s docs page you can find a guide howto use RapidMiner with those sandbox environment.

Use the Server!

With RapidMiner 7.2. the server also got a free version. It has it’s computational limitations — 2GB of RAM, one logical processor — but it is still a nice piece of software to use. You can now set up your server to collaborate with your colleagues. You can operationalize your processes using the web services, webapps or cron jobs.
I use a server also as a local computation machine. It runs in the background and simply does it’s job. Just have a look on the docs page for a installation guide.

Customize your RapidMiner to your needs

RapidMiner is a versatile piece of software and can be aligned to your needs. With a fresh install you might not have all the Panels you might want to have. You can activate new Panels under View->Show Panel.
My must haves are:

  • Context
  • Macros
  • XML
  • Server Monitor

and for the oldschool of us: The Tree View (it was called View back then!)

Here is my current setup:

Get into Gradient Boosting

RapidMiner includes now a parallized version of Gradient Boosted Trees! You might want to say: “Meh, this is the 1001 learning method in RapidMiner!” and you are technically right, but Gradient Boosting had extreme success in recent data mining competetions. It is one of the strongest algorithms around. I’ve used this implementation in our internal beta phase and I am amazed by it.

To understand the algorithm just have a look at this fantastic guide to gradient boosting regression trees Thanks to AYLIEN for sharing this guide on Twitter!

Get Into Deep Learning

Besides Boosted Trees RapidMiner 7.2 also features a Deep Learning operator! Deep Learning is the most hyped technique in machine learning for quite some time.

If you are used to RapidMiner already the new Deep Learning operator is not be that different to you. In fact the good “old” Neural Net operator is very similar. The new Deep Learning operator adds a lot of algorithmical and numiercal twists to it. You can add Regularization, change activation functions or include drop out, but the general idea is similar.

If you are new to Neural Nets and Deep Learning, have a look on this article from kdnuggets.

Read The Release Notes

Each RapidMiner release has a huge set of small enhancements which might be beneficial for your analysis. I encourage you to have a look at the release notes to be aware of what else changed!

Share your experience in the RapidMiner Community

Data Science is a team sport. You cannot win without working together. So have a look at the RM community and join it!