machine_learningヘッダファイルパッケージで用いている計算式

1. 数学関数の定義と微分形

関数sigmoid

(Formula used in machine_learning header file package; 1. Definitions and differential forms of mathematical functions — function sigmoid)



◆定義(Definition)

\[\begin{equation} f(y)=\frac{1}{1+\exp(-y)} \end{equation}\]

◆微分形(Derivative)

\[\begin{eqnarray} f’(y) &=& -\frac{-\exp(-y)}{[1+\exp(-y)]^2} \nonumber \\ &=& \frac{\exp(-y)}{[1+\exp(-y)]^2} \nonumber \\ &=& \frac{1}{1+\exp(-y)}\frac{\exp(-y)}{1+\exp(-y)} \nonumber \\ &=& \frac{1}{1+\exp(-y)}\frac{[1+\exp(-y)]-1}{1+\exp(-y)} \nonumber \\ &=& \frac{1}{1+\exp(-y)}\left[1-\frac{1}{1+\exp(-y)}\right] \nonumber \\ &=& f(y)\left[1-f(y)\right] \end{eqnarray}\]