bias-variance tradeoff

The bias-variance tradeoff is a human-made concept in machine learning describing the fundamental tension between two competing sources of prediction error: bias, which is the error introduced by approximating a real-world problem with a simplified model (causing underfitting), and variance, which is the error introduced by the model's sensitivity to small fluctuations in the training data (causing overfitting). Total prediction error decomposes as bias squared plus variance plus irreducible noise. The tradeoff posits that model complexity increases variance while decreasing bias, creating an optimal complexity point where total error is minimized. This concept persists through mathematical frameworks (error decomposition theorems), empirical visualization (learning curves, validation plots), and practical methodologies (ensemble methods like bagging reduce variance, boosting reduces bias). [formal: commercium bias-variantiae | substrate: mind | horizon: a life | explicit: yes | epoch: 1.00]