Repeatability

Repeatability, in the context of Machine Learning, is the closeness of model training and prediction results such as:

• Accuracy

• Bias

• Confidence Levels and Intervals

• Precision

• Receiver Operating Characteristic

• Sensitivity/Recall

• Variance

Repeatability is enhanced when Modeling Processes and Data inputs are kept constant.

Even when these are kept constant, there will be some variance due to probabilistic uncertainty. An example of this is the Confidence Intervals and Levels related to measured values:

References

• Web Search

• Wikipedia Article

• YouTube Search