Gradient Boosting

Fact

An ensemble method that adds weak models to fix earlier mistakes.

In Plain Words

Gradient Boosting is like a team of tiny tutors. Each new tutor works on yesterday’s wrong answers.

You meet it in spreadsheets for yes-or-no calls and number guesses. It keeps fixing earlier errors, round by round.

Related Concepts

• Gradient Descent
• Classification Task
• Regression
• Bias-Variance Tradeoff

Gradient Descent
Gradient Boosting uses Gradient Descent to choose the next fix.

Classification
Gradient Boosting is often used for spam checks and risk decisions.

Regression
Gradient Boosting also predicts numbers, like home prices or sales.

Bias-Variance Tradeoff
Too many deep trees can lower bias but raise variance.