Chapter 1: Models and Cost Function

• Supervised Learning: A model whereby you are given the “right answer” for each example in the training data.

• There are two types of supervised learning problem: regression and classification.

• In a regression problem, you want to predict a real-valued output.

• The other type of supervised learning model is called classification, where the aim is to predict discrete-valued output.

• In regression, the hypothesis refers to when you are mapping x features to y predictions.

• A linear regression model with one variable (x) is also know as a univariate linear regression.

• Cost function let’s the machine figure out how to fit the possible line through our data. What values of the parameters gives the minimum error.

• The squared error cost function is one of the most commonly used cost functions for regression problems. There are other ways to calculate the objective cost function as well.

• Contour plots are great to visualize the cost function since using the original hypothesis and it’s parameters will generate a bow shape 3D figure that is difficult to read.