RapidMiner 9.8 Beta is now available

Be one of the first to get your hands on the new features. More details and downloads here:

GET RAPIDMINER 9.8 BETA

Moving average for each ID

Cristina_daimielCristina_daimiel Member Posts: 2 Newbie
Hello all,

I have a dataset with the energy produced by several PV plants each 15 minutes across 1 year. Therefore, I have a column with the datetime (around 18000 examples for each ID), another one with the ID (each PV plant have a different ID, in total I have 4 IDs) and the energy produced. For each example, I'm calculating the moving average of the previous 3 hours with the Operator "Moving average filter". However, when the first year of the first ID ends, for the second ID, the moving average is is calculating the average for the last 3 hours of the previous ID, instead of starting the calculation from the beginning. Is there a way for me to take the ID into account to do this calculation? Or should I separate the exampleset into 4 different exampleset (one for each ID) and do the calculation separately?

Many thanks in advance
Tagged:

Best Answer

Answers

  • Cristina_daimielCristina_daimiel Member Posts: 2 Newbie
    mschmitz ! It's working now

    After the loop Connection operator, the example set have been split into 4 different dataset (one per PV plant) within an IOObjectcollection. Do you happen to know how can I combine again the data into the same exampleset?

    The project I'm working on has as its objective the prediction of failures in a photovoltaic plant. For this I have data from different variables, together with enviroment conditions (irradiation, temperature, humidity, etc) every 15 minutes and for a full year. 
  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,636  RM Data Scientist
    the append operator is able to append horizontally back to one example set. I recommend to use the operator Append (Superset) which has some more advanced features, but the normal append should so do it.

    Let me know if we can help you further. Especially if this is a commerical opportunity.

    Cheers,
    Martin
    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
Sign In or Register to comment.