Is standard deviation reported when the Loop Parameters operator is used?

caesar

Hi. I have a CV where, if run on it's own (outside of a Loop Parameters operator) I get average accuracy and

the standard deviation. However, when I run my CV within the Loop Parameters operator, changing the Neural Net training cycles, I only get accuracy reported for each of my training cycle settings. Is it possible to have the standard deviation returned also - is there some setting I must apply?

Thank you.

YYH

Hi @caesar,

You are able to get the standard deviation of the performance metrics from Cross Validation. 

The log will record the average performance from the cross validated models. To get more details, you may have to convert performance to data (example set) and then post-process tables for a summary ("append" for all iterations).

Attached is an example process. Hope it helps.

Cheers,
YY

caesar

Thanks very much, that does the job, appreciate that.

YYH

Hi @caesar,

You are able to get the standard deviation of the performance metrics from Cross Validation. 

The log will record the average performance from the cross validated models. To get more details, you may have to convert performance to data (example set) and then post-process tables for a summary ("append" for all iterations).

Attached is an example process. Hope it helps.

Cheers,
YY

caesar

Thanks very much, that does the job, appreciate that.