"I want to use only when my model have high accuracy"
HI!
I use Rapidminer Studio 6
I am developing some accounting model for predicting posting Account for transactions. it is confidential so I don't give more info but the scenario doesn't matter b/c my question is general.
I used x-validation and I got 87 % accuracy as I can see from the PerformanceVector table. I want to use the model but since the application area is sensitive I want to use it only
when it has high accuracy.
what I want :
I give unseen example to the model then :
- the model predicts if its 97% sure ( accurate)
- if not it says I don't know
I hope my question make sense
I use Rapidminer Studio 6
I am developing some accounting model for predicting posting Account for transactions. it is confidential so I don't give more info but the scenario doesn't matter b/c my question is general.
I used x-validation and I got 87 % accuracy as I can see from the PerformanceVector table. I want to use the model but since the application area is sensitive I want to use it only
when it has high accuracy.
what I want :
I give unseen example to the model then :
- the model predicts if its 97% sure ( accurate)
- if not it says I don't know
I hope my question make sense
Find more posts tagged with
Sort by:
1 - 1 of
11
I found the answer ... there is already an operator that to exactly this "Drop Uncertain Predictions". 