Isn't there a weight given with the predictions given by the Classifier
output?
In the multistepBestPreductions dictionary you have {Probability:
prediction, .. }

Yiu can use the probability to prioritize.
I think that is what the demos do as well...
I'd have to check though, I'm on the road so no computer available ;)
 On Jul 28, 2015 1:47 AM, "Chandan Maruthi" <[email protected]>
wrote:

> I have a question on predictions. Any insight into this will be great.
> I am training a region with multiple different sets of data.
> For instance
>
> Dataset 1
> 1
> 2
> 3
> 4
> 5
>
> Dataset 2
> 2
> 4
> 6
> 8
> 10
>
> Dataset 3
> 1
> 3
> 5
> 7
> 9
>
> Now when I try to predict for 2, I expect to see 3 and 4, Without any
> additional data, which I do. Is any information available on the strength
> of the prediction, that can be used to influence the way the prediction is
> used.
>
>
> --
> Regards
> Chandan Maruthi
>
>

Reply via email to