Need help choosing two time series courses

[mathtoquant]

There are two courses that I can take related to time-series in UChicago

1. Financial Statistics: Time-Series, Forecasting, Mean Reversion, and High Frequency Data (from FinMath department)
2. High Dimensional Time Series Analysis (from Statistics department)

For quant research role, would #2 substitute #1?
I wanted to take #1 but it turns out that I have no more room to take it that particular semester. So I am now thinking of taking #2. Could anyone give me advice for me?

 

[AyeRasputinTheTasteful]

Well, it seems like your hands are tied, no? I think the High Dimensional Time Series course could be beneficial. There are many areas in Finance, especially in practice, where dimensionality causes issues (e.g. yield curve analysis, copula modelling). Learning the basic tools to compensate for that (PCA/SVD/Factor Models) will certainly be useful. That is, if those are topics covered in the course. Do you have a syllabus or brief description of the topics that would be covered? I would imagine there would still be some overlap with the other course: vector autoregression models, ARCH/GARCH models, etc.

Obviously, a time series course from the Financial Math department will be more tailored to your wants... but if there aren't other courses you're interested in for that semester, the time series course from the Statistics department will at least introduce you to a set of tools you can apply to finance.

 

[mathtoquant]

Well, it seems like your hands are tied, no? I think the High Dimensional Time Series course could be beneficial. There are many areas in Finance, especially in practice, where dimensionality causes issues (e.g. yield curve analysis, copula modelling). Learning the basic tools to compensate for that (PCA/SVD/Factor Models) will certainly be useful. That is, if those are topics covered in the course. Do you have a syllabus or brief description of the topics that would be covered? I would imagine there would still be some overlap with the other course: vector autoregression models, ARCH/GARCH models, etc.

Obviously, a time series course from the Financial Math department will be more tailored to your wants... but if there aren't other courses you're interested in for that semester, the time series course from the Statistics department will at least introduce you to a set of tools you can apply to finance.

Thank you for your insight!
I can't access to the syllabus, but I have the following description for the course! Do you think it will be useful? I feel like this could be more advanced and need some preparation

*High Dimensional Time Series Analysis
This course will include lectures on the following topics: review of asymptotics for low dimensional time series analysis (linear and nonlinear processes; nonparametric methods; spectral and time domain approaches); covariance, precision, and spectral density matrix estimation for high dimensional time series; factor models; estimation of high dimensional vector autoregressive processes; prediction; and high dimensional central limit theorems under dependence.

 

[AyeRasputinTheTasteful]

Do you have any background in basic time series analysis? Does the course have any listed prerequisites? I think a brush-up on elementary time series topics (notion of stationarity/heteroskedasticity/unit roots/cointegration, AR/MA/ARIMA models) would probably be beneficial prior to starting the course. But in the end, I still do think it would be useful. Are there any other viable courses you're interested in for the semester?

 

[mathtoquant]

Do you have any background in basic time series analysis? Does the course have any listed prerequisites? I think a brush-up on elementary time series topics (notion of stationarity/heteroskedasticity/unit roots/cointegration, AR/MA/ARIMA models) would probably be beneficial prior to starting the course. But in the end, I still do think it would be useful. Are there any other viable courses you're interested in for the semester?

Thanks for the suggestion! No I do not have any background in it, but I have relatively heavy pure math background and enters as a graduate student, so hopefully they will be lenient for prerequisites. I will take this course next year, so I think I have enough time to prepare as well!
Another option I have is Monte Carlo Simulation from Stat department and "Advanced computing for finance" from FinMath program.
I think both will be very practical and useful but I feel like time series course should be my priority.

 

[AyeRasputinTheTasteful]

Hmm... yeah, all three courses actually sound pretty good if you ask me. Knowing Monte Carlo methods is big; anecdotally, I use these almost every day. "Advanced computing for finance" sounds like a programming course with a finance bias? Again, would be very useful too... you need to know how to code in Python/C++/C# for any kind of quant position. But that is something you could probably pick up yourself relatively easily - compared to the other options. As a corollary, getting familiar (if you're not already) with proper coding style guides is a good habit to pick up - and not something that's as ubiquitous in the industry as it should be. See, e.g., styleguide

Anyway, you probably won't go too far wrong with the time series course from the Stats department. If that falls out, though, I would recommend the Monte Carlo course.

 

[mathtoquant]

Hmm... yeah, all three courses actually sound pretty good if you ask me. Knowing Monte Carlo methods is big; anecdotally, I use these almost every day. "Advanced computing for finance" sounds like a programming course with a finance bias? Again, would be very useful too... you need to know how to code in Python/C++/C# for any kind of quant position. But that is something you could probably pick up yourself relatively easily - compared to the other options. As a corollary, getting familiar (if you're not already) with proper coding style guides is a good habit to pick up - and not something that's as ubiquitous in the industry as it should be. See, e.g., styleguide

Anyway, you probably won't go too far wrong with the time series course from the Stats department. If that falls out, though, I would recommend the Monte Carlo course.

Awesome. Thanks so much!