Data Science: Regression Modelling
Tutors: Maxwell Munford & Morgan Nightingale (GTAs)
Course Level: Level 2
Course Credit: 1 credit
Prerequisites: Knowledge of basic statistical concepts. Basic knowledge of MATLAB would be beneficial.
Duration: 3 hours
Format: Live online or live face to face with hands-on practice.
This workshop will combine an informal lecture and practical modelling to explain how to apply regression methods to model data in terms of one or multiple variables. We will discuss how to approach modelling problems and draw conclusions from correlated variables. This workshop will focus on the application and methods for using regression, NOT the theory, and will include a practical session followed by how to interpret and analyse your model. The practical session will require very basic level use of MATLAB, but all necessary scripts will be provided.
- steps in regression modelling process
- what do we assume?
- data exploration
- example - car emissions data
- regression - residuals and least squares
- regression with a single predictor variable
- multiple explanatory variables
- binary and categorical variables
- confounding factors
- model interactions
- model checking
- model fitting
- hands-on example
- explore a model created using MATLAB (code will be supplied)
- tune the model
On completion of this workshop you will be able to:
- Identify the correlation coefficient as a single measure of linear association.
- Apply linear regression to model a response variable in terms of a single or multiple variables
- Assess model validity by checking model assumptions.
- Assess model fitness by comparing the results produced by the model with your data.
Dates & Booking Information
- Wednesday 15 December 2021, 10:00-13:00, MS Teams
- Wednesday 26 January 2022, 10:00-13:00, MS Teams
- Wednesday 16 March 2022, 10:00-13:00, MS Teams
- Wednesday 27 April 2022, 10:00-13:00, MS Teams
- Wednesday 29 June 2022, 10:00-13:00, MS Teams
Please select a date and book on via Inkpath using your Imperial Single-Sign-On.