Demystifying Machine Learning

Summary:

• Loss function/Cost function: your explicit definition of what you mean by the ‘best’ model.
• A common loss function for linear regression: sum of squared errors (across all training examples)
• The objective of your algorithm is to minimise the loss function.