Describe the bug
When a classification model is run, a predicitons.csv file is not generated. This is critical for our ontology. While the predicitions.csv file will naturally be different for classification than regression. But one option on how to approach this is to make 0=incorrect and 1=correct, then we take the average of correct vs incorrect. For multi-label, we can do the same thing, but maybe in a matrix. Then the overall error could be the average of the matrix.
Describe the bug When a classification model is run, a predicitons.csv file is not generated. This is critical for our ontology. While the predicitions.csv file will naturally be different for classification than regression. But one option on how to approach this is to make 0=incorrect and 1=correct, then we take the average of correct vs incorrect. For multi-label, we can do the same thing, but maybe in a matrix. Then the overall error could be the average of the matrix.
Expected Outcome
Current Outcome
Platform, Desktop