Measuring the bias of a machine learning model


How can we measure the bias of a machine learning model? Can we determine it by just calculating its performance estimates difference on the train data and test data? For example, if a model SVM predicts on train data is 0.53 and then on test data it is 0.60; can we say that the bias is just 7 points?

I have read an article which mentioned that "The bias of a model validation technique is often measured in terms of the difference between a performance estimate that is derived from a model validation technique and the model performance on unseen data."



Posted 2020-04-13T15:16:50.417

Reputation: 11

No answers