Model accuracy is a measure of how well the model fits the data, specifically how close the predicted response is to the true response.
Accuracy is often measured in terms of [[mean squared error]] for [[regression]] models.
In [[classification]] problems, [[error rate]] is typically used.