"Random forest: Show attribute weighting for each examples"

Hello.

The idea is that I want to create a ranking of attributes that positively influenced the Confidence(class A) for each examples in my testset (customers in this case).

So, I'm wondering if there is a way to get the attribute weights (Weight by Tree Importance, random forest) on the example level in the testset (instead of the attribute weights of the model overall).

I think that it should be possible but I coundn't find an operator to do it...

Thank you for you help.

Cheers