# GLM or Statistical Inference for Machine Learning?

First of all: I don't know if this is the correct place to ask this question. If it is not, please help me deciding where I should post it.

I want to study Machine Learning as a MSc (UCL if possible) and my background is undergraduate mathematics. I'm in my second year and in my third year I can choose only one of the following two statistic modules:

• Statistical Inference(Point Estimation: MLE, MVUE. Hypothesis Testing: Neyman-Pearson Lemma, Likelihood ratio test. Computational inference: MLE, Resampling methods. Bayesian Inference: prior proba, post prob, predictive inference. Decision Based Inference)
• Generalized Linear Model (syllabus here)

what is the best choice? Which one is more useful in ML?

edit The courses I've done so far are:

Y1: Calculus, Multivariable Calculus, Linear Algebra 1, Linear Algebra 2, Mathematical Modelling, Intro Stats and Probability, Differential Equations, Number Theory&Cryptography.

Y2: Statistical Distribution Theory, Intro Applied Mathematics, Analysis, Vector Calculus&Complex Variables, Statistical Methods 1 (linear regression), Applications of vector calculus (tensor, fluidodynamics, electromagnetism, etc), Algorithms (trees, flows), Parial Differential Equations.

• What is otherwise your interests? Which other courses did you take? Commented Mar 25, 2017 at 14:57
• @kjetilbhalvorsen I've edited my question adding the courses that I've taken! Commented Mar 25, 2017 at 15:08
• Both not very helpful for traditional machine learning. Commented Mar 25, 2017 at 15:50
• @StudentT see that's the problem. There is so much contraddictory information on the web. For example a lot of posts say that statistics and maths in general are essential and the most part of ML. But then others say it's not. So why this difference? Is stats used in ML? Commented Mar 25, 2017 at 20:12
• Sorry to add to the confusion, but i think both those are essential for machine learning :) Commented Mar 25, 2017 at 20:34

It really depends on your interests and must be up to you. Sooner or later you will need a course in inference. But at this stage, it might be better to go for GLM course first, if its a good one (like the one I would like to give now, not the one I actually gave years ago ...). But you did'nt state the syllabus of the GLM course.

For some information and opinion about what could be a good GLM course, see Reference Request: Generalized Linear Models and http://www.math.unm.edu/~bedrick/glm/glm.html

• I'll state the syllabus soon! Also, I'll put the link of the Msc I would like to do (Computational Statistics and Machine Learning at UCL). If I post the link to the course, which shows modules as well and their syllabuses, would you be able to then tell me (knowing what to expect in that course) the best choice between those 2 modules? Commented Mar 25, 2017 at 20:15
• I've put the syllabus on the question. You can see it here: southampton.ac.uk/maths/undergraduate/modules/… Commented Mar 25, 2017 at 20:45