Jump to content
Urch Forums

babyrudin

Members
  • Posts

    18
  • Joined

Converted

  • My Tests
    No

babyrudin's Achievements

Newbie

Newbie (1/14)

5

Reputation

  1. If you've finished Spivak, then just read the first 4 or 5 chapters of Lay. That would be more than enough to start Rudin. It's not as scary as some people say. S&B can be useful, but given your background (calc 1-3, lin alg, diff eqns), there's nothing much S&B can teach you in terms of math (of course, you might still want to see applications to economics), and it definitely won't help much in terms of preparing for Rudin.
  2. Everyone stresses the steeper learning curve of actual typesetting in LaTeX but forgets to mention the tremendous speed that you can achieve once you get used to it. One can now honestly type mathematics much faster than actually handwriting it. This simply cannot be achieved using, say, Scientific Word or Lyx (imagine trying to type out six lines of first order conditions using Microsoft Equation; that's basically the speed you can achieve using Scientific Word or Lyx).
  3. Learn from the master: Mattuck's lectures on differential equations.
  4. Are you looking for something like this: http://www.albany.edu/econ/eco_phds.html
  5. I started a page on mathematics, mostly to set up a structure that I think may be suitable for economics (list out the math topics, and state in which areas of economics they are used). I found myself in the awkward position of saying that __________ is useful in every branch of economics for all the three topics that I wrote about. But that's just because I was writing about calculus, analysis and real analysis, and I expect that problem to be less pervasive for other branches of math. I didn't want to spend too much time on it so there's little on the page right now. But I guess that's the advantage of this Wiki thing; others will pick it up and write (or rewrite) more.
  6. It's clear, intuitive, but not rigorous. Math in PhD programs are way beyond that, though to learn those you probably need to know the stuff in Chiang anyway.
  7. Generally speaking both are well-written though I personally prefer Spivak because it is SO well motivated. Note that Apostol treats integration before differentiation and some find that awkward. I wouldn't worry too much about Vol 2 of Apostol - most likely your course will not get to Chapter 9 of Rudin, which is what Vol 2 of Apostol is about. As a bridge for Rudin, however, I don't think those two are the best choices. In my experience most students have trouble with Rudin NOT because it is hard, but usually because they are not familiar with techniques of proof. Hence if this is your first proof-based course the priority is to learn the techniques of proof. In this regard, both Spivak and Apostal are not suitable. The preferable choice is one that both treats the techniques of proof and analysis together so you'll get practice in both areas. There are many such books available now. A popular one is Lay's Analysis with an introduction to proof. Also don't worry too much about reviewing Stewart or Simmons. Knowing how to integrate trigonometric functions and the like is nice but will not help at all in terms of understanding Rudin.
  8. The difficulty of MWG is overrated. I also don't understand why some say that MWG is dry. The motivation and commentary provided are numerous and well-written. Instead of worrying about MWG, I would worry more about Stokey-Lucas. That's if you are going to take PhD Macro of course.
  9. I think you have misplaced your topic as Excel should have nothing to do with PhD in economics. The PhD's in business should be more able than us to answer your question: http://www.www.urch.com/forums/phd-business/
  10. If you want to do serious econometrics or macro (heck, even micro), you should use Linux, so that when you're estimating a model that takes 3 weeks to run it won't crash on the 20th day. . . . As for software, Matlab is a general numerical one that most economists are at least somewhat familiar with, even if that's not what they use. Try to avoid loops (vectorize) if possible; it's quite sluggish at that. Matlab can do algebraic manipulations now, but learning Maple may also be desirable as the latter is arguably the best at playing with algebra. Gauss is also a general numerical one, usually faster, but less extensive with inferior graphical capabilities. Octave is a Matlab clone that is free. Ox is another numerical package, and it's probably by far the fastest of all the ones mentioned. Anyhow, you will probably need to learn at least one of these for macro. For econometrics, any of the above can be used as well as long as you're willing to program yourself, though most people use more specialized econometrics/statistic packages. Stata is popular both in the industry and academia. It has unmatched panel data capabilities. TSP (Time Series Package) is popular among older generation econometricians. Its nonlinear estimation is more reliable than any other I've tried, mostly because of its use of analytic (first) derivatives. Shazam is a well-rounded one, but is horrible at nonlinear anything, unless they have improved dramatically recently. R is very popular among statisticians. It is free and its add-on packages are ever expanding thanks to contributors worldwide. I also found its syntax to be especially intuitive. R has a commercial version called S-Plus, which has nicer interface and tech support. Unfortunately, both R and S-Plus are very slow with loops, so you want to avoid them as much as possible. The industry also likes SAS, which is actually used more as a data management package nowadays rather than econometric analysis. One thing worth mentioning about econometrics/statistics software is that don't be surprised if they do things incorrectly. In fact, I have come to the conclusion that as a general principle, econometrics packages cannot be trusted. One should never use a preprogrammed function unless either you know how it was written (unlikely) or it provides good, clear documentation on what exactly it is doing (which is, unfortunately, unlikely as well). Currently I have reverted back to Matlab, though a switch to Ox might be coming. By the way, Matlab (or Gauss or any matrix language) is extremely useful for learning econometrics. The value-added from writing your own OLS, GLS, IV, MLE, ..., estimators should not be underestimated. In this case you should also be familiar with a general purpose econometrics software, so you can compare and check you results. Of course, if you are not trained in numerical computer science, most of what you write will run very inefficiently, which gives you motivation for also learning numerical analysis, something any good econometrician should be very familiar with. If you know lots about numerical methods, well, that's motivation for putting them into good use! . . . Another thing I would just like to add that I found recently is that R also has very well-written random-number generators. Probably again due to its connection with statisticians.
  11. I don't think people really choose an operating system simply based on availability of software. Anyhow, all of the software you mentioned except EViews and SPSS run on Unix/Linux. EViews is most popular among undergrad students as it is dead easy (mostly "point-and-click") and good as an introduction to statistics/econometrics software. SPSS is popular among psychologists and has great graphics, but is not powerful enough otherwise for econometricians. I wrote a review on another topic about economics computing and econometrics software more specifically awhile ago. I guess it's more suitable for this topic so I will just repost it.
  12. I haven't heard of any recent placements into top US programs, but I believe they occasionally place into schools like UWash.
  13. SFU Master graduates, having done well in the program, should have no problem at all getting into other PhD programs in Canada.
  14. McGill at least hire some of its PhD graduates. SFU never. SFU PhD graduates usually work at community colleges. Probably because of the location, McGill also does substantially better in getting their grads into institutions like Bank Of Canada, Department of Finance, Competitive Bureau, etc.
  15. In order of whichever comes to mind first. Economics: Econometrica, American Economic Review, Journal of Political Economy, Journal of Economic Theory, Quarterly Journal of Economics It is hard for a field of economics to have 5 "top" journals, but I guess you're asking for top 5, so here's my list: Finance: Journal of Financial Economics, Journal of Finance, Journal of Banking and Finance, Journal of Business, Journal of Financial and Quantitative Analysis
×
×
  • Create New...