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

This is the time series Problem. Where 26 weeks data is provided and we ave to predict next 4 weeks

pallavpallav Member Posts: 38 Contributor II
edited March 26 in Help
In the given problem we have to forecast next 4 weeks of data give past history of past 26 weeks for each client and each product .
We have to forecast  product wise for each client .so ideally in the data four more column(week27,week28,week29,week30) will be added with the forecast result. Attaching dummy data in attachment.

How to do this problem in rapid miner, I tried doing it in rapidminer but was not able to do it . It will be help full  if i can get a demo model of this. I will learn from it and try to implement my problem in rapidminer.
Tagged:

Best Answer

Answers

  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,622  RM Data Scientist
    Hi @pallav ,
    the key operator is called "Multi Horizon Forecast". It allows you to build n-models to predict the n next points. Attached is an example on the german gas station data set. It forecasts the next 6hours of gas station price.
    Best,
    Martin


    <?xml version="1.0" encoding="UTF-8"?><process version="9.6.000">


    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
  • pallavpallav Member Posts: 38 Contributor II
    @mschmitz- I cannot find  attached example 
  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,622  RM Data Scientist
    Hi,
    the XML is in the spoiler.Just click on "show".
    Cheers,
    Martin
    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
  • pallavpallav Member Posts: 38 Contributor II
    @mschmitz - unfortunately i can see any process inside it 
    One one line is there.

Sign In or Register to comment.