Forecasting a price with and without the influence of a specific input variable

Melvin_pslMelvin_psl Member Posts: 1 Newbie
Hello, I would like to forecast short-term electricity prices (intraday electricity prices) for the German electricity market and measure the influence of the input variable "control energy quantity". I have several input variables (wind energy, solar energy, balancing energy quantities,...). The price I want to forecast is the intraday price. I have built the following process (see screenshot). This gives me a forecast for intraday prices. Now I want to measure the influence of the input variable of the control energy quantity. So I want to measure the prices once with the input variable and once without the addition of the variable. To compare the forecast prices with and without the influence of the balancing energy quantity. I think that would work with weights? Do you have any ideas on how I can build the process? 


  • Options
    MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,525 RM Data Scientist
    you can just set the role of the undesired feature to metadata right before your NN. It will not use the model then.

    I would recommend to use the Deep Learning operator instead of the NN operator, since it is more versatile. It also provides feature weights.

    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
Sign In or Register to comment.