Linear Regression Models · Statistics
I learnt how to apply linear regression models to solve real life problems.
Lots of proofs and, by virtue of doing so many proofs, some proof techniques. Ordinary least squares regression, generalized least squares, weighted least squares, logistic, penalized. Diagnostic testing, back testing.
Homework in R (Not weekly- Due about every 3rd class, but again, soft deadlines, especially if he felt the class was struggling on particular concepts. We only actually got through 7 of them, and 8 was posted but not graded over reading week.) Midterm 1: Half T/F, Half short answer problems. No direct coding problems (almost entirely conceptual with a few minor computations in short answer). Average around 85. Midterm 2: This one was all short answer/ short computations. I think average was still around 85. Final: A LOT of T/F ( around 75%) and only 2 short answer. If you've done every HW yourself and read the book well, you should understand the concepts enough to do fine on T/F.
I learn Linear Regression Models in this class!
I learned concepts on multiple forms of linear regression (simple linear regression, linear regression with multiple predictors, Lasso, Ridge, Logistic...) and observed the derivations of their formulas. I also learned how F-tests and T-tests are applied in linear regression.
I learned I could pay college tuition to have a professor read me a textbook going to this class is like buying an e-book on Amazon
Linear Regression - Simple Linear Regression, Multiple Linear Regression, Matrix Formulation, Interpreting/Understanding Summary Statistics, Implementation in R
Simple linear models. Multiple linear models. ANOVA. Hypothesis testing ...
~5 problem sets. Each was similar in length to what I've had in other classes with 10 or 11 problem sets, but twice the time. Midterm and final. Both were fairly long, not tooooo tricky but still somehow difficult. Mostly because he gave us no preparation for them.
Weekly problem sets -- usually 3-6 question with a combination of proofs, theory questions, and computational problems. That's 40% of the grade + midterm (20%) + final (40%).