Supervised Learning

Given a training set, we want to get a learning algorithm so that our hypothesis function $h(x)$, can be a good predictor for the corresponding values of $y$.

Linear Regression

When the target we are trying to predict is continuous, it is a regression problem. The most popular of which is Linear Regression

Classification

When the target is discreet, it is a Classification problem.
We will take what we learned from Linear Regression and adapt it to our Classification problem.
We do this by taking the values that we would normally have as large values in our parameter $\theta$ and changing them to fit into discreet values, where $y \in {0,1 ...}$
This can be done by the Sigmoid Function