0 Introduction
1 Supervised learning
- supervised learning：have training set. given the ‘right answer’ for each example in the data
1.1 Regression problem
- regression problem：predict real-valued（实数值） output
1.2 Classification issues
- classification problem:predict discrete-valued（离散值） output
2 Unsupervised learning
Notation
- m m m —— Number of training examples
- x ′ s x's x′s —— ‘input’ variable/features
- y ′ s y's and′s —— ‘output’ variable/‘target’ variable
- (x, y) (x, y) (x,y) ——one training example
- ( x ( i ) , y ( i ) ) (x^{(i)},y^{(i)}) (x(i),and(i)) —— i t h i^{th} ith training example
3 Fitting
3.1 Underfitting
- underfit (underfit) / high bias (high bias)
3.2 Overfitting
- overfit (overfitting) / high vorionce (high variance)
- if we have too many features, the learned hypothesis may fit the training set very well（ J ( θ ) ≈ 0 J(\theta)≈0 J ( θ )≈0）, but fail to generalize（泛化） to new examples
4 Reference
Wu Enda machine learning coursera machine learning
Huang Haiguang machine learning notes