Classification methods results through Testing and validation
Hi
I wonder how I can get the accuracy, recall, precision, f-measure of the classifier from testing and validation?. I have multi class classification. I used cross validation. I got the final accurcy, recall , precision of the model, but I want to know how was model doing in training and testing.
Regards
Muhanad
I wonder how I can get the accuracy, recall, precision, f-measure of the classifier from testing and validation?. I have multi class classification. I used cross validation. I got the final accurcy, recall , precision of the model, but I want to know how was model doing in training and testing.
Regards
Muhanad
Find more posts tagged with
Sort by:
1 - 2 of
21
Typically you shouldn't care about the training error. It is always overfit and that is why it is not a reliable measure of performance in any way. The cross-validation will deliver the testing error, which is what you should care about.
If you insist on getting the training error for whatever reason, then you will have to build a process to capture the model performance on the training side (inside cross validation) and then use the Log operator to keep that information.
If you insist on getting the training error for whatever reason, then you will have to build a process to capture the model performance on the training side (inside cross validation) and then use the Log operator to keep that information.
To have a general idea of the training performance, you can put Apply Model and Performance operators
in the training part of the Cross Validation operator and then connect the PER output port to the THR port.
Here a such process :
Lionel