Machine Learning

Ensemble Methods

Manipulation approaches for Ensemble Learning

Ensemble Methods

Stacking

Simple voting

Meta classification

Nested cross-validation

Assessment

Bagging

Assessment

Random Forest

Random tree

Random forest

Assessment

Boosting

AdaBoost

Output: weighted set of base classifiers: ${(\alpha_1, C_1), …, (\alpha_T, C_T)}$

Computing $\alpha$

alpha

Updating $w$

\[w_j^{i+1} = \frac{w_j^i}{Z_i}\exp{-\alpha_i}\] \[w_j^{i+1} = \frac{w_j^i}{Z_i}\exp{\alpha_i}\]

weight update

Classification

\[C^{*}(x) = \argmax_y \sum_{j=1}^{T}\alpha_j\delta(C_j(x)=y)\]

Assessment

Bagging vs Boosting

Bagging

Boosting


Edit this page.