IITM BS Linear Models with Applications (BSMA3015): Syllabus and Tips

Linear Models with Applications (BSMA3015) is a 4 credit degree level elective on building and checking linear models with real data. The first 8 weeks cover least squares, model checks, transformations, logistic regression and generalised linear models (GLMs). Weeks 9 to 12 are a guided project with a written report. The course page lists no prerequisites. The instructor is Siva Athreya, who also teaches Linear Statistical Models.

Not in the handbook's course table

The DS handbook's degree level table (updated 18 March 2026) does not list this course, though it lists its sister course Linear Statistical Models (BSMA3012). So the handbook gives no tag or term plan for BSMA3015. The details above come from the course page. Under the handbook's rule, a 3xxx code means a level 3 course.

What you learn

• Weeks 1 to 3: the linear model. A review of estimation, hypothesis testing and R, plus the project handout. Then least squares, estimable functions, normal equations, indicator variables and interaction terms, BLUEs and the Gauss-Markov theorem.
• Weeks 4 and 5: checking and fixing a model. Model adequacy checks, residual analysis and residual plots, points that pull the fit too much, transformations that make a model linear, variance-stabilising transformations, the Box-Cox transformation, and generalised and weighted least squares.
• Weeks 6 to 8: beyond ordinary regression. Logistic regression, GLMs for the exponential family with their log-likelihood, gradient and Hessian, the distribution of the MLE, prediction, testing linear hypotheses, and choosing a model with AIC and BIC.
• Weeks 9 to 12: the guided project. You carry out the project and write the report. The project is discussed in week 5, and you choose it in week 7.

How it is assessed

The course page lists 12 weeks of coursework, weekly online assignments, 2 in-person invigilated quizzes and 1 in-person invigilated end term exam. The week list also includes the guided project, but the page does not say how the project is graded or how much it counts. Check the grading document for your term.

Where it counts

• Degree level elective. The course page lists it as an elective worth 4 credits.
• BSc elective credits. As a 3xxx course it should count as level 3, and level 3 credits can fill the 8 BSc elective credits. Confirm with support, since it is not in the handbook table. See the BSc degree level.
• No minor. It is not in any minor list.

This course or Linear Statistical Models?

The two share their early topics. Pick based on what you want.

The handbook does not say whether both can count toward your degree. Ask support before you take the second one. More in the Linear Statistical Models guide.

Who finds it hard and how to prepare

The course packs a lot of theory into 8 weeks, and then the project takes over. If you fall behind in the first half, you end up catching up on theory while the project runs. Reading residual plots is also a skill you build only by seeing many of them.

• Pick data early. You choose the project in week 7, so have two or three data sets in mind by week 5.
• Revise Stats 2 and Maths 2. Estimation and testing come up in week 1. The Hessian from Maths 2 comes back in week 6 for GLMs.
• Practise in R. Fit a linear model and a logistic regression, then plot the residuals. Do this on small data before the term.
• Keep a project log. For each model you try, note why you tried it and what the residuals showed. Most of your report can come straight from this log.

What to take before and after

• Before: Maths 2 and Stats 2. If you chose Option 1 in the Diploma in Data Science, Business Analytics already covered regression basics, diagnostics and logistic regression. This course goes deeper.
• After: Statistical Computing shows how models without a closed form solution are fitted, with Newton's method, gradient descent and EM.
• Compare options: the DS electives list.

Common questions

Is the guided project a separate project course?

No. The project sits inside the course weeks, not as a separate project course like the diploma projects. How it is marked is not stated on the page, so read the grading document.

When is this course offered?

The handbook's term plan does not include it. Look for it in the course list at registration, or ask support.

Can I show the guided project in my portfolio?

A finished analysis with a clear report can be a good portfolio piece. Check the course rules on sharing first, and never share graded work in a way that helps others copy it. See building a portfolio from your projects.