Hot to select attributes based on aggregated weights?

nejckozar

I would like to make a selection of let's say top 10 attributes selected by using aggregation of more than one weighting operators and apply this selection to original example set.

Select by weight operator can use only weight input and there is no operator for weight aggregation.

 

I can only use weight to data and perform aggregation afterwards, then use sorting and filtering, but then is is impossible to use this data again as weights. (data to weights will always produce 1)

 

Is there any other way to achieve this selection?

 

Thanks, 

 

Nejc

lionelderkrikor

Hi @nejckozar

 

Hi the data to weights operator can not perform your task. Maybe it is an improvement for a next release of Rapidminer...

 

I think there is a "trick" adapted to your problem with the Write Weight Values operator and  Read Weight File operator to convert

data into weight. It is described in a previous topic :  

 

https://community.rapidminer.com/t5/RapidMiner-Studio-Forum/Setting-Custom-Weights-From-Example-Set-Table/td-p/11148

 

I hope this will be helpful.

 

Regards,

 

Lionel