qqyyzz

Congrats on such nice placements. I think if you're even remotely interested in machine learning or data mining go with CMU by far.

I have a bad feeling since Michigan and NC State still haven't responded to me yet.

Well from what I understand it's a number of reasons. First, suppose you go to one school for a terminal masters then a different school for PhD. A lot of schools don't let you transfer course work and you'll end up being forced to take a lot of overlap courses when you enroll in your PhD program. Second, if you decide to do masters but you really want a PhD keep in mind that PhD programs are likely funded whereas masters programs are likely not funded. Third, a lot of schools mention that their master and PhD programs are separate; finishing masters somewhere doesn't mean they're more likely to accept you into their PhD. I can't decide whether it's better to go for a master's at a higher ranked school or a PhD from a lower ranked program. But that's just a few things to keep in mind. Btw, I was wondering if you guys that were accepted could provide your profiles seeing as how you were in one of the first waves of acceptances for UC Davis and I'm still waiting here sitting on my hands

If you're interested in a PhD, I think you should probably go for a PhD directly. Professors I've talked to have advised against going to terminal masters then PhD if you're just interested in PhD from the start. With that said, I applied to davis and haven't heard anything yet

As promised, I have some early results from some schools: CMU joint statistics/machine learning: Rejected Minnesota: Waitlisted Wisconsin: Accepted with TA/RA position Ohio State: Accepted, waiting for letter detailing offer Still waiting to hear back from a ton of schools. Good luck to everyone else this cycle!

Please tell me the iowa state and NCSU ones are fake too

I applied to Duke too, since I haven't gotten a response I assume that's not good . What are your grades, research experience and background like if you don't mind my asking?

Oh, I am also applying to chapel hill (UNC). I'm mostly interested in machine learning and data mining which I did mention in my SOP. I mostly followed the NRC rankings when I applied although I did take a look at some of the US news and world report rankings. Fortunately many of the 30-50 ranked schools have later deadlines (feb/march).

Oh I should be a bit clear, the school itself isn't unknown, but more specifically the math phd/ms program here is unranked. I am interested in machine learning hence the comp sci courses. My research is in the econ department but it is basically mathematical modelling, more math than econ. I do wish I had taken more stats courses, but honestly the pickings are pretty limited here. I'll post results here when I get them in a few months.

Already applied to a bunch of places that had dec 15 deadlines: NCSU, UNC, Duke, UCLA, CMU, WashU, UC Irvine, Minnesota and Wisconsin. Also applied to Rice and Purdue. Unfortunately, Purdue is one of the schools that recommends the math subject GRE which I have not taken.

Now that I finished most of my apps for this cycle (except for some of the later due date schools), I was wondering if you guys could evaluate my chances. I know it's a bit unusual to evaluate one's profile after they have already applied, but I didn't want to get discouraged from applying to any schools I was interested in and wanted to remain optimistic. Well I guess I'm ready to take a beating now. Graduated economics BA from small liberal arts school GPA 3.28 (3.33 major) - pretty crappy here. Went back to state university to take more courses a couple years later, here are the grades for my math courses at state university (relatively few math courses from previous university): Calc I, II, III - A, B, A Lin Alg. - A+ ODE - A+ Intro proof course - A Intro Probability - A+ Numerical Analysis I, II - both A+ Intro to real analysis - A Statistical Inference - A Discrete Math - A+ Abstract Alg. - A+ Upper level Probability - A+ Advanced Calculus - A+ Pattern Recognition (grad level Comp Sci course) - A Will take in spring: Algorithms, Advanced Calc II, upper division linear algebra, data storage. GRE: 800 math, 630 verb, 5.0 writing Worked as a math tutor and a grader for linear algebra for one year. Research assistant for economics department in game theory (this project was funded by a grant since this fall). Worked unpaid at this position all last year, but fall 2011 was funded by grant and I was paid as a grad level research assistant. Used many computer models for this project. Concerns: did pretty bad for my initial undergrad degree with almost no math courses (got a C+ on calc I and a withdraw in calc II when I was at liberal arts univ). Recommendations are from two professors in the math department and one from econ. The math department is theoretical in nature and these professors are probably not well known in statistics circles. Additionally, this state university is not the highest in rankings when it comes to math (in fact it is unranked according to US news).

I'm currently an unclassified graduate student at a state university (just taking math courses basically) and one of the professors I've been researching for just got a grant and offered me a tuition waiver for next year. Unfortunately due to university policy I am unable to take advantage of this due to being an unclassified student. I've talked to the graduate director of the math department and he says I can still apply to the master's math program for next fall (very soon I know). The problem is the following: I was planning on taking math courses in the fall along with applying for statistics programs (my interest is in statistics and I was hoping to apply for stats programs next fall for the 2012 year). I was planning on taking these math courses as an unclassified student and paying out of pocket. Aforementioned math courses could also count toward the math master's and I could essentially get a year of tuition and save myself money. The problem is that I would enter this master's program and leave after just one year. Is it worth it for me to do this for a tuition waiver? When I do apply for statistics programs in the fall will this potentially reflect poorly? Some basic background information: my interest is in statistics and many of the math courses I've been taking are oriented toward getting into a good stats program (math stats, analysis etc). The courses I had planned on taking next fall include more analysis and probability theory. The university is a state university with an unranked math phd program. There is no stats or applied math program at this school and the department is overwhelmingly theoretical pure math. The professor who is offering the tuition waiver is actually not in the math department but in a closely related field (due to my background and the nature of research the math program was a likely choice).

The subject test is a beast. Based on the information you listed, about 30-40% of the content on the math GRE is going to be stuff you've never seen before. If you are still set on taking it, besides calculus, and linear algebra you should probably learn abstract algebra, some real analysis, discrete math and a little bit of general topology.

Hi Christina, I can't speak for HCI but if you're interested in the machine learning side of AI you will need to take a fair amount of mathematics. You mentioned you had taken statistics, but for a graduate degree in this area they would probably expect you to have a calculus based probability/stats class and probably a discrete math course. Also, you mentioned algebra in college/high school, was this course in elementary algebra (variables, arithmetic, polynomials), linear algebra (matrices, vector spaces, linear transformations) or abstract algebra (groups, rings)? You will need to be comfortable in linear algebra; the concept of vector spaces is used in many places and many things will be written in matrix form. Don't worry though, if you don't have these things, you can always take a few courses before you apply. I was an econ major and subsequently found out I was more interested in data mining/large data set problems and have decided to take a bunch of math courses so that I am ready to apply next fall for stats programs.

This is all anecdotes, but from what I've heard many of the people studying machine learning come from math backgrounds and many places will look very specifically at math background for machine learning. Unfortunately, every CS program I have looked at requires (obviously) a bunch of CS courses as admission requirement. I'm not sure how people without a CS major get into these programs; I assume that many of these people majored in something else but have close to the number of CS courses as a major. Alternatively, you could look for statistics phd programs that do research in machine learning/statistical learning. After all, machine learning is a field that's at the intersection of computer science and statistics and many of the machine learning algorithms are the same as statistics techniques with different terms.

What is your background in math? This might be bad news but the math GRE is a tough test and it will be even tougher if you haven't taken some of the topics. Yes you can self study, but there will be some topics that are difficult to self study. Analysis, for example, is difficult to self study as a lot of books don't have answers for proofs exercises. As a result you might go about a lot of proofs the wrong way and have no one to tell you. Besides that I think at a very minimum for math phd programs you are going to need all of calculus, linear algebra, abstract algebra, real analysis and/or advanced calculus, and perhaps complex analysis. Keep in mind that's probably an absolute minimum. Realistically math programs would probably expect at least 30 credit hours of upper division math courses. You might be able to study discrete math and combinators in the CS department if that's what you're interested in. Although I expect that probably depends on which CS department you are at.

Thanks for detailed post Kash. I'm actually in a pretty similar situation as the OP. I graduated three years ago from a small liberal arts school in econ. Unlike the OP however, my GPA was not as good and I only took two math classes that I did horrible in (C+ in calc I and withdraw from calc II). However, I've since retaken calc I online from UC berkeley with an A and enrolled in my state university (not the best university overall, but it's the biggest in the state and decently cheap with in-state tuition). And I've taken the rest of calc series, ODEs, numerical analysis, prob theory, linear algebra, an intro proof class with mostly As. This semester I'm taking real analysis, numerical analysis II, and math stats. I plan on taking abstract algebra, discrete math, advanced calculus, PDEs and advanced prob theory before I apply next fall. I knew coming from a non-quantitative background and with mediocre grades this would be an uphill battle. But I'm glad to hear that the adcoms don't look at non-math classes too heavily.

Thanks for the detailed answer! That helps me out a lot.

I'm interested in econ stats. However, I'm just curious about the differences as I've posted above. Additionally, I noted that someone said to avoid biostats for reasons listed above and I'm wondering if the same reasoning applies to econometrics.

Hi everyone, I'm trying my best to prepare for graduate school. Initially I was interested in economics, more specifically econometrics. However, I've been doing a bit of research and I found that econometrics and statistics have a lot of similarities. What exactly are the differences between PhD level theoretical econometrics and statistics? I've been searching the web for awhile and can't find any definitive answer. I've been reading through this forum and there was a thread a couple of months back about statistics PhD programs. One of the posters, who had been admitted to a PhD program, warns against going into biostatistics. His reasoning was almost any biostats PhD job can be done with a regular stats PhD and biostats PhDs are less likely to find employment in finance and insurance etc. Since there are obvious similarities between econometrics and statistics, does this line of reasoning also apply to econometrics? Can statisticians work most econometric jobs? Can econometricians do a statistician's job? What courses should I take to maximize my chance of getting into a statistics phd program? In the course for preparing for an econ PhD, I have already taken/plan on taking the following: calc through multivariate, linear algebra, differential equations (ordinary and partial), real analysis, probability theory and statistical inference. My university does not have any grad classes in stats, but would a grad class in econometrics help in that area? Finally, this summer and next semester I have the opportunity to do graduate level game theory research (math and programming intensive). I know this would help me for econ grad school, how much would it help with a statistics phd program?