What Is the Bias-Variance Tradeoff?
The bias-variance tradeoff describes the balance between systematic model error and sensitivity to changes in training data.
Don't Just Read it – Explore It
The bias-variance tradeoff describes the balance between systematic model error and sensitivity to changes in training data.
Training, validation, and test sets divide data into separate roles for fitting, selecting, and finally evaluating a model.
Underfitting happens when a model is too limited or insufficiently trained to capture important patterns in its data.
Naive Bayes classifies examples with Bayes’ theorem while assuming features are conditionally independent given the class.
Logistic regression is a linear classification model that converts a weighted feature score into a class probability.
A nearest neighbors method predicts from the labeled training examples that are closest to a new case under a chosen distance.
Boosting builds learners sequentially so later models concentrate on examples or residual errors the current ensemble handles poorly.
Gradient boosting adds weak models stage by stage to reduce a differentiable loss, often using decision trees as the learners.
Bagging trains multiple versions of a model on bootstrap samples and combines their predictions to reduce variance.
Weak supervision creates approximate training labels from rules, heuristics, distant sources, models, or other noisy signals.
Ensemble learning combines predictions from multiple models to improve accuracy, stability, or robustness.
Online learning updates a model incrementally as examples or small batches arrive instead of fitting once on a fixed dataset.