吴恩达机器学习第三章总结

逻辑回归

logistic regression: $0\le h_{\theta }(x)\le 1$

$$h_{\theta }(x)=\frac{1}{1+e^{-{\theta }^{T}X}}$$

decision boundary(决策边界)

${\theta }^{T}X>0$ predict y=1;
是假设本身及其参数的属性所决定的

$$J(\theta )=\frac{1}{m}\sum _{i=1}^m(-y^{(i)}log(h_{\theta }(x^{(i)})-(1-y)log(1-h_{\theta }(x^{(i)}))$$

advanced optimization(高级优化)

options = optimset(‘GradObj’, ‘on’, ‘MaxIter’, 100);
initialTheta = zeros(2,1);
[optTheta, functionVal, exitFlag] = fminunc(@costFunction, initialTheta, options);

muticlass classification

$$h_{\theta }^{(i)}(x)=P(y=i|x;\theta )(i=1,2,3)$$

over fitting

regularization(正规化）
underfit high bias
overfit

1.reduce number of features
2.regularization

$$J(\theta )=\frac{1}{2m}\left[\sum _{i=1}^m(h_{\theta }(x^{(i)})-y^{(i)}{)}^2+\lambda \sum _{j=1}^n{\theta }_j^2\right]$$