Machine Learning Evaluation Metrics
Evaluation metrics help to evaluate the performance of the machine learning model. They are an important step in the training pipeline to validate a model. Before getting deeper into definitions and types of metrics, we need to understand what type of machine learning problem we are solving. Classification metrics differ from regression metrics. These metrics influence how we weight the importance of different characteristics in the results and our ultimate choice of which algorithm/model-version to choose.
Classification metrics
- Accuracy
- Precision
- Recall
- F-Score
- ROC (Receiver operating characteristic)
- AUC (Area under the curve)
Let’s have a look at the listed metrics above one by one.
Accuracy:
The first metric to evaluate when it comes to a classification problem is the accuracy, which can be calculated using the confusion matrix.
This is how a confusion matrix looks like.