Week 2

This week we review Machine Learning Fundamentals.

Resources

Lecture	Slides
1	Slides

Topics

Prediction

The purpose of Machine Learning Models is to predict new outcomes (\(Y\)) from given data. This is achieved my constructing the function \(f\) that predicts \(Y\):

\[ Y = f(X;\theta) \]

where \(X\) is the new data obtained and \(\theta\) are the parameters that dictate how a predefined model¹ \(f\) behaves.

For the purposes of this class, there are two types of predictions: Regressiona and Classification.

Regression

The

Classification

Types of Learning

Supervised Learning

Unsupervised Learning

Model Adequecy

Regression Loss Function

Classification Error Rate

Machine Learning Training Fundamentals

Train-Test Split Data

Standardization of Data

Pipeline Coding

Footnotes

1. There are models that relax this assumption.