Skip to main content

MH3510 Regression Analysis

Course Summary​

This course aims to study the statistical relationship between predictor and response variables. The course covers content including:

  1. Simple Linear Regression
  2. SLR with Matrices
  3. Multiple Linear Regression
  4. Polynomial Regression
  5. One/Two-Way Classification
  6. Analysis of Variance (ANOVA)
  7. Analysis of Categorical Variance
  8. Model Selection

The course starts off by teaching Simple Linear Regression, which is a topic that everyone should be familliar with given that it should be covered in Introduction to Data Science & AI. It makes use of concepts learnt in Statistics such as parameter estimation, confidence intervals and levels of significance.

Workload​

The content does take some getting used to but the structure of the course is cyclic. I.e., for every topic topic taught, the prof first dives into the explaining the different terms of the model, then he will explain the calculations to get the various coefficients, how to write out the ANOVA table, and finally how to test if the model or its component variables are significant.

The lectures are 4 hours per week and are inclusive of a 1 hour tutorial conducted during the lecture itself. Tutorial worksheets contain at most 10 questions and are very straightforward to do, albeit a bit tedious at times depending on the calculations. There are also a couple of tests and one Final constituting 50% of the grade.

Projects​

There is one group project to be done in R and submitted together with a report around Week 10.

tip

The project is done in group sizes of 10 and he provides a sample from the previous year's project as a reference. Most of the project is modifying the sample task to fit the current data.

Things to take note of​

  1. He can throw curveballs in the final exam and test obscure parts of the content so do be prepared for that.