What is bias in machine learning models and how is it introduced?

Bias is the error introduced by simplifying assumptions in the model.

Bias refers to systematic error introduced when a model is too simple to capture underlying patterns in the data. High bias leads to underfitting. It is introduced by overly constrained models (like linear regression on nonlinear data), insufficient features, or strong regularization. Bias is one component of the bias-variance tradeoff and directly affects training error.