Noesis
2.3.11 · HinglishTree-Based & Instance Methods

AdaBoost algorithm

301 words1 min read↑ Read in English

2.3.11 · AI-ML › Tree-Based & Instance Methods

Weak learner kya hota hai?

Final AdaBoost classifier ek weighted vote hai

Figure — AdaBoost algorithm

Concept Map

combines

built

start with

measures

sets

used to

focus on

derived from

combined into

grows on mistakes

contrasts with

AdaBoost = Adaptive Boosting

Weak Learners = stumps depth 1

Trained sequentially

Sample weights w_i

Weighted error e_m

Vote power alpha_m

Re-weight data

Weighted vote H of x

Exponential loss e^-yf

Overfits noisy data

Bagging = parallel