Chapter 17 Logistic Regression

17.1 Related concepts

        Logistic regression is more about modeling from a probability level, but in practical applications, it is more used in binary classification problems.

        Thinking: How to calculate "better fitting", analogous to the least squares method of linear regression.

        sigmoid function:

y = \frac{1}{1+e^{-x}}

\frac{dy}{dx}=\frac{d[\frac{1}{1+e^{-x}}]}{dx}=\frac{d(1+e^{-x})^{-1}}{dx}

=-(1+e^{-x})^{-2}\times(e^{-x})

Guess you like

Origin blog.csdn.net/qq_36171491/article/details/124860597