Parameter Optimization in Rapidminer 5.0

In several of my video tutorials I assign different parameters for my learning model on the fly.” Of course the question any astute reader/viewer should ask is, why did you choose those parameters instead of another combination?”

That’s a great question and the answer is, well I just choose those parameters to illustrate my point for the video.” While this answer is not at all satisfying to the astute reader/viewer, it does lead us to ask the most important question of all, what are the right parameters to choose?”

This can be  answered very well if you were to use Rapidminer’s Parameter Optimization operator in your initial data discovery phase. This operator allows you to choose some or all of the parameters in your experiment and iterate different values for them to meet some specific requirement on your part (i.e. performance).

For example, if you were using the Neural Net operator and didn’t know what to set your learning and momentum parameters to, to get the best classification accuracy, you would use the Parameter Optimization operator to interate different combinations of those parameters to find the best accuracy.

Once the Parameter  Optimization  operator determines those values, you can input  them into your experiment and truly optimize your model for performance! See below for an actual output from a parameter  optimization  model I’m working on. You can see that Rapidminer indicated that momentum of 0.3 and a learning rate of 0.6 was the best parameter settings to maximize the accuracy rate and minimize the classification error.

While is operator is a fantastic feature (they got evolutionary optimizers too!) for us data modelers, its a massive computer resource suck.  I would advise anyone using this operator to have a very powerful server or computer, with oodles of memory, to run your iterations.


9 thoughts on “Parameter Optimization in Rapidminer 5.0”

  1. Congrats on the impressive performance.

    What data did you use? and what are the inputs?

    Also are you finding Neural Nets to be more accurate then SVMs?


    1. @Cleo: It depends on the task at hand, in some cases NN’s work better than SVMs. You can let the parameter optimization figure out which one is best, but it requires time and computer power.

  2. Hello Tom,

    Thanks for the response. My setup does has a select subprocess operator which contains a few different svm, and nn configurations but as you said it takes up alot of computer power.

    I am curious about your inputs? Could you please let me know what features you used, ie moving average, open, close price, etc… and for which instrument and timeframe?

    Thanks again for the videos, and i hope you enjoy your summer vacation.


  3. Helo Tom,
    I’ve learned statistic, I’ve studied your post, but my
    question is, why did you chose this kind of output variable?

    Is it useful, for predict the future? for recognize the
    current trend? or?

    I mean is there any sense to predict (estimate) the
    current state of the price? (DOWN, UP, SIDEWAY)
    or the next one?

    Could we create a profitable strategy with this kind of

    How long should I hold a long position after an UP

    Or I’ve missed something, that is why I wrote my
    questions, (( not for attack your great and helpful

    // ps.: sorry for my English (I am Hungarian)

    Best wishes
    Istvan Pintye

  4. @Istvan: The UP/DOWN/SIDEWAY tags I assigned based on a visual inspection of the time series data. I was interested to see what the trends break so I looked at the peaks and valley’s of the price action.

  5. Hello sir, can you explain a tutorial regarding the application of the
    algorithm Particle Swarm Optimization in RapidMiner, please.

    1. Hi Juliano! That’s a great idea and a cool tutorial, but it won’t happen anytime soon because of my workload (I can barely keep up posting here!). Maybe in March? Maybe. 🙂

      1. Ohh.. sorry sir 🙂 I guess you have leisure time. But if you can, i hope can be learning sir. :)Thanks for your other tutorials.

Comments are closed.