Machine Learning Week Two

notation

$x^{(i)}_j$denotes value of feature j in the ith training example

$n$ denotes the number of features

The multivariable form of the hypothesis function is

$$h_\theta(x)=\theta_0+\theta_1x_1+...+\theta_nx_n$$
or

$$h_\theta(x) = \sum_{i = 0}^{n}\theta_ix_i$$ ’
where we asume $x^{(i)}_0=1$
we alse write $$\theta=\begin{bmatrix}\theta_0\\\theta_1\\\vdots\\\theta_n \end{bmatrix}$$
and
$$X = \begin{bmatrix}x_0\\x_1\\\vdots\\x_n \end{bmatrix}$$
so we can rewrite h(x) as
$$h_\theta(x) = \vec\theta^T\vec X$$

Algorithm

$$\theta_j:=\theta_j - \alpha\frac{1}{m}(h_\theta(x^{(i)}) - y^{(i)}) )x^{(i)}_j$$
(for j = 0 to n)
where x0 = 1

Algorithm in practise

feature scaling

We can speed up gradient descent by having each of our input values in roughly the same range.

$$x_i:=\frac{x_i-\mu_i}{s_i}$$
where $\mu_i$ means the average of xi
and $s_i$ means the range of xi or means standard deviation

learning rate

Debugging gradient descent.

Make a plot with number of iterations on the x-axis. Plot the cost function, J(θ) over the number of iterations of gradient descent. If J(θ) ever increases, α is too large.

Automatic convergence test.

Declare convergence if J(θ) decreases by less than E in one iteration, where E is some small value such as 10−3. However in practice it’s difficult to choose this threshold value.

Features and Polynomial Regression

we can combine multiple features into one. Such as x3 := x1 * x2
We can change the behavior or curve of our hypothesis function by making it a quadratic, cubic or square root function (or any other form)for example:
$$h_\theta(x) = \theta_0+\theta_1x_1+\theta_2x_1^2+\theta_3x_1^3$$
or
$$h_\theta(x) = \theta_0+\theta_1x_1+\theta_2\sqrt x_1$$

Normal Equation

$$\theta=(X^TX)^{-1}X^T\vec y$$
where
$$X=\begin{bmatrix}x^{(1)^T}\\x^{(2)^T}\\\vdots\\x^{(m)^T} \end{bmatrix}_{m\times (n+1)}$$
and
$$y = \begin{bmatrix}y^{(0)}\\\vdots\\y^{(m)}\end{bmatrix}$$
differences

Gradient Descent	Normal Equation
Need to choose alpha	No need to choose alpha
Needs many iterations	No need to iterate
$O(Kn^2)$	$O(n^3)$ and need to calculate X’X
work well when n is large	Slow if n is very large