使用培训和Sli窗口时间滞后ding Window Validation Operator

pix123

Hi there,

I am new to rapidminer and having issues in running a model to predict appliance usage based on a time series. The time is in 10 minute intervals. There are some additional pre-processing steps also included.

The process is failing to run and I can't figure out how to fix or improve. Would welcome any suggestions.

Thanks.

Current process...

<运营商激活= " true " class = " filter_examples”compatibility="8.1.001" expanded="true" height="103" name="Filter Examples" width="90" x="447" y="34">





<运营商激活= " true " class = " set_role”兼容ibility="8.1.001" expanded="true" height="82" name="Set Role" width="90" x="581" y="34">

lionelderkrikor

Hi@pix123,

Can you share your dataset in order we can reproduce the error ?

Regards,

Lionel

pix123

Thanks Lionel, please find attached.

lionelderkrikor

Hi again@pix123,

Here you can find a process to forecast "Appliances".

1. In this version of process, Window size =20, you can modify this setting by :

- setting this parameter inside theSet Predictions_Paramssubprocess operator.

and

- modifying the parameters ofSelect AttributesandRenameoperators inside theLoopoperator.

2. To improve this process, you can test differents models (Deep learningmodel is chosen arbitrarily....)

and you can use theOptimize Parametersoperator to determine the combinaison of parameters to obtain the best performance for your model.

The process :

futureMonths
15


horizon
1


windowSize
6
































<参数键= value = " RH_8.true.real.attribu“18”te"/>




<参数键= " 23 " value = " RH_out.true.real.attribute"/>



































<运营商激活= " true " class = " set_role”兼容ibility="5.3.013" expanded="true" height="82" name="Set Role" width="90" x="447" y="136">









<运营商激活= " true " class = " filter_examples”compatibility="6.4.000" expanded="true" height="103" name="Filter Examples" width="90" x="849" y="85">













































































<运营商激活= " true " class = " set_role”兼容ibility="5.3.013" expanded="true" height="82" name="Set Role (2)" width="90" x="447" y="165">

I hope it helps,

Regards,

Lionel

pix123

Hi@lionelderkrikorapologies for the delay. Thank you for this, it has been helpful.

My machine currently has performance issues due to the memory needed to run a large process.

Is there an efficient way to run a prediction model without the need to window?

Thanks again.