Description

Multifeature linear regression with,

• Feature scaling
• Vectorized implementation

Dataset

UCI Machine Learning Repository - Boston Housing

Model

$w_{1}x_{1}+w_{2}x_{2}+...+w_{13}x_{13}+b$

Cost

(Not exactly right because of multiple weights but gives a good enough idea) $$\frac{1}{2m}\sum_{i=1}^{m}(f(x^{(i)})-y^{(i)})^2$$

Derivatives for gradient descent

For setting weight $w_j$

$$\frac{1}{m}\sum_{i=1}^{m}(f(x^{(i)})-y^{(i)})x^{(i)}_{j}$$

Constant $b$

$$\frac{1}{m}\sum_{i=1}^{m}(f(x^{(i)})-y^{(i)})$$

Output screenshot