CORE Skills Data Science Springboard - Day 5 - Simple predictions, regression and statistical model building
The aim of today's session will be to introduce simple linear regression as the basis for prediction, and discuss the factors that alter its accuracy and effectiveness. We'll move onto more complex linear regression situations and show how even some non-linear datasets can be used with linear regression, and then we'll show how these tools can also be used classification problems. We will make our first machine learning models in this session.
You should aim to understand the basics of regression and outliers, how regression models can be limited by statistical assumptions about the data and how to recognise when these assumptions are being violated. You'll should also understand the principle of Occam's Razor and how to choose between basic statistical models while evaluating their effectiveness.
While linear regression is something you've most certainly encountered before, we'll be framing much of today's work in the language and approaches of machine learning, which will be applicable over the next few weeks.
Todo: there is no pre-reading at this time. Consider browsing through the program which has links to further resources to browse.