I'm working on a few RNN (Recurrent Neural Network) models and want to evaluate those models, so I'm looking for useful metrics to evaluate RNN models?
1
-
2$\begingroup$ What would you use for a different kind of model? $\endgroup$– DaveFeb 26, 2022 at 3:56
Add a comment
|
3 Answers
$\begingroup$
$\endgroup$
Frame challenge: The metric you use to evaluate your models should be based on the task you're trying to solve and things like distribution of labels in the data, not the type of model. As long as the results of the different models all have the same form e.g. a real-valued vector, you can use the same metric to evaluate them regardless of model type.
$\begingroup$
$\endgroup$
Evaluation metrics are dependent on the machine learning task you are performing.
This can be classification (typical metrics are precision, recall, AUC, F1, etc.), regression (MSE, MAPE, ...), or something else (e.g., for image segmentation you can use intersection-over-union).
So, first define your task, and then look at what metrics are used for that task - the actual algorithm / model you're using is irrelevant for determining the evaluation metric.
$\begingroup$
$\endgroup$
You can use Mean Squared Error or Root MSE. With hyper tuning, try to reduce the values to the minimum
