Variance
Variance is a statistical measure that describes how far individual values in a dataset tend to differ from the mean. In machine learning, it is also used to describe a model's sensitivity to changes in its training data.
What is Variance?
Variance measures the average squared difference between individual values and their mean. In the context of machine learning, high model variance means a model may fit its training data very closely and produce substantially different predictions when trained on different datasets.
Why is Variance Important?
Understanding variance helps evaluate how consistently a model performs and whether it may be overfitting. A model with excessively high variance can perform very well on training data but poorly on unseen data. Variance is a key component of the bias-variance tradeoff.
Common use cases
Variance is commonly used in statistical analysis, model evaluation, uncertainty analysis, feature analysis, and understanding overfitting in machine learning.