mdabros/SharpLearning

Understand class prediction results

rick779 opened this issue · 2 comments

rick779 commented

Hi,
maybe some silly questions...
I'm trying to train a random forest model with two different classes. I think I understood that the number of rows of the target vector must be equal to observations matrix (therefore regardless of the number of classes). So in the rows of the output vector I set the value 0 for the first class and 1 for the second class. This is right?
I would also like to understand how to interpret the results, for example if for a set of features I have the prediction value 0.6 I must consider it a class of type "0" or a class of type "1"? Do I have to cast to integer or I must to round it? Finally the "variance" values ​​contained in CertaintyPrediction indicates the probability of the prediction (greater is better)?
many thanks

mdabros commented

Hi @rick779,
Since you closed the issue again I assume you found answers to your questions?
best regards
Mads

rick779 commented

Yep, for the classification I was using RegressionRandomForestLearner when I realized that there is ClassificationRandomForestLearner.
thank you, great project

  • 👍1