Spring Courses...advice

[hedgie]

I am looking to do a PhD in Stats. I have limited time in the spring but could take a one or both: a stochastic processes course, and numerical linear algebra. Any advice whether these would be of value, or could one just self study the stochastic?

[cyberwulf]

Stochastic might expose you to a few more topics you'll cover in grad school, but I would view both as "mathematical culture" courses, to take if you find them interesting.

[hedgie]

What exactly do you mean by "mathematical culture" courses? I assume these will only be helpful in certain areas of stats that I do not know if I will go into or not. I know Stochastic is good for time series but who knows what I'll do. I guess my best option is wait till grad school and take them if I need them or read the books and learn it myself.

Thanks for any comments thoughts and your initial response!

[cyberwulf]

"Mathematical culture course" = "course where you'll be exposed to more mathematical ideas, which is generally a good thing if you're going to be studying statistics."

[Poisson]

If you have the time, I would recommend taking both classes. In most graduate programs, you will cover some, but probably not all material from both courses (e.g. linear models will probably cover Cholesky/QR decompositions methods to solve the OLS equations) so having the background will be very useful.

I took stochastic processes before starting grad school, and I found it very useful as I ran into Markov Chains and Poisson Processes in several classes, and it definitely helped to have the background. I wish I could have taken numerical linear algebra, but it was relatively easy to self-study (probably easier than self studying stochastic processes). If you have the time, I would also highly recommend taking measure theory if you plan on entering a more theoretical program.

[hedgie]

Thanks so much cyberwulf and Poisson!

When you say numerical was easy to pickup on your own do you just mean a intro to numerical analysis or would the numerical linear algebra be similar and easy to study on my own?

I wish we had a measure theory course offered but alas my school does not and I can it step into the real analysis 2 grad course...so I'll have to wait until next year.

[Poisson]

When you say numerical was easy to pickup on your own do you just mean a intro to numerical analysis or would the numerical linear algebra be similar and easy to study on my own?

I mainly looked through the relevant wikipedia articles (Cholesky, QR, Singular Value Decomposition, Principal Components, etc), and used this book http://www.amazon.com/Numerical-Analysis-Statisticians-Statistics-Computing/dp/1441959440/ref=sr_1_1?ie=UTF8&qid=1325446448&sr=8-1 as an additional reference (concentrating on the numerical linear algebra chapters).

[hedgie]

Thanks Poisson!