李航统计学习方法－－定理 8.1（AdaBoost 的训练误差界）

skmygdrs

已于 2022-05-20 14:59:27 修改

阅读量158

点赞数

文章标签：深度学习机器学习人工智能

于 2022-05-19 16:29:56 首次发布

本文链接：https://blog.csdn.net/skmygdrs/article/details/124863146

版权

$e_m=\sum_{i=1}^N{w_{mi}I\left( G_m\left( x_i \right) \ne y_i \right)}\,\,\left( 8.1 \right) \\ \alpha _m=\frac{1}{2}\log \frac{1-e_m}{e_m}\,\,\left( 8.2 \right) \\ Z_{mi}=w_{mi}\exp \left( -\alpha _my_iG_m\left( x_i \right) \right) \\ Z_m=\sum_{i=1}^N{Z_{mi}}\left( 8.5 \right) \\ w_{m+1,i\,\,}=\frac{Z_{mi}}{Z_m}\left( 8.4 \right) \\ f\left( x \right) =\sum_{m=1}^M{\alpha _mG_m\left( x \right)}\left( 8.6 \right) \\ G\left( x \right) =sign\left( \sum_{m=1}^M{\alpha _mG_m\left( x \right)} \right) \left( 8.7 \right) \\ \frac{1}{N}\sum_{i=1}^N{I\left( G\left( x_i \right) \ne y_i \right) \leqslant \frac{1}{N}\sum_{i=1}^N{\exp \left( -y_if\left( x_i \right) \right)}}=\prod_{m=1}^M{Z_m}\left( 8.9 \right)$

$\sum_{i=1}^N{I\left( G\left( x_i \right) \ne y_i \right)} \\ =\sum_{G\left( x_i \right) =y_i}{0}+\sum_{G\left( x_i \right) \ne y_i}{1} \\ =\sum_{y_if\left( x_i \right) >0}{0}+\sum_{y_if\left( x_i \right) <0}{1} \\ \leqslant \sum_{y_if\left( x_i \right) >0}{\exp \left( -y_if\left( x_i \right) \right)}+\sum_{y_if\left( x_i \right) <0}{\exp \left( -y_if\left( x_i \right) \right)} \\ =\sum_{i=1}^N{\exp \left( -y_if\left( x_i \right) \right)}$

$\frac{1}{N}\sum_{i=1}^N{\exp \left( -y_if\left( x_i \right) \right)} \\ =\frac{1}{N}\sum_{i=1}^N{\begin{array}{c} \exp \left( -y_i\sum_{m=1}^M{\alpha _mG_m\left( x_i \right)} \right)\\ \end{array}} \\ =\sum_{i=1}^N{w_{1i}\prod_{m=1}^M{\exp \left( -y_i\alpha _mG_m\left( x_i \right) \right)}} \\ =\sum_{i=1}^N{Z_{1i}\prod_{m=2}^M{\exp \left( -y_i\alpha _mG_m\left( x_i \right) \right)}} \\ =\sum_{i=1}^N{Z_1w_{2i}\prod_{m=2}^M{\exp \left( -y_i\alpha _mG_m\left( x_i \right) \right)}} \\ =\sum_{i=1}^N{Z_1Z_{2i}\prod_{m=3}^M{\exp \left( -y_i\alpha _mG_m\left( x_i \right) \right)}} \\ =Z_1\sum_{i=1}^N{Z_{2i}\prod_{m=3}^M{\exp \left( -y_i\alpha _mG_m\left( x_i \right) \right)}} \\ =\prod_{m=1}^{M-1}{Z_m}\sum_{i=1}^N{Z_{Mi}} \\ =\prod_{m=1}^{M-1}{Z_m}Z_M \\ =\prod_{m=1}^M{Z_m}$