What is bias in cost function optimization?
Updated May 15, 2026
Short answer
Bias refers to systematic error in model predictions.
Deep explanation
High bias models underfit data, leading to consistently high cost even on training data.
Real-world example
Using linear model for nonlinear data.
Common mistakes
- Confusing bias with bias term in neural networks.
Follow-up questions
- How to reduce bias?
- Bias vs variance tradeoff?