Stats PhD, Masters - Profile Evaluation

[engineered_features]

School: Large Public Research University (40-45 ranked math department - USnews ranking)

Major: Mathematics, Computer Science

Minor: Economics

GPA (Cumulative): 3.81/4.00 (Math): 3.9

Mathematics Coursework: Calculus I,II,II (A-,B,A-), Discrete Math (Intro to proofs) (B+), Linear Algebra (Proof-Based) (A), Ordinary Differential Equations (A),  Real Analysis I (A-), Real Analysis II (A), Abstract Algebra I (A-), Complex Analysis (A), Linear and Nonlinear Optimization (A), Financial Mathematics (pretty much just a course in stochastic calculus) (A), Numerical Analysis (next year), Partial Differential Equations (next year)

Statistics/CS Coursework: Probability Theory (A), Statistical Theory (A), Graduate Machine Learning (A), Data Mining (A), Financial Econometrics (A), Time Series (next year)

Math Courses with Ws (and then retook for A/A-): Calculus III, Abstract Algebra I, Probability Theory

GRE Scores: 169Q, 165V, 4AW (I was going to apply this year, but decided to finish up my computer science degree by taking a 5th year, and add some more research to my profile)

Letters: 2 from professors having done research, 1 from professor I took several classes with (opportunity to do research with next year too - see below)

Research: -non-relevant to statistics (both in number theory): Semester Research Project, Summer REU

                 -relevant: year long  research opportunity (if I accept - see below)

 

PhD Programs (preferred) : Chicago, Michigan, Harvard, CMU (joint machine learning - honestly, I never see anyone get admitted to this though), Berkeley, Wisconsin, Penn, Penn St, Duke, University of North Carolina, North Carolina State, Columbia, Cornell, Purdue, UCLA

MS Programs: Harvard, Stanford, Chicago, CMU, Berkeley, Duke, Washington, UCLA (trying to keep this top heavy, since the consensus is that masters are easier to be admitted to)

 

I have quite a bit of interest in machine learning, and would prefer going to a program that has some faculty doing research in this area, but I wouldn't say it is a deal breaker (i.e. if I was admitted to Chicago I'd go, regardless of the minimal research in ML). 

 

Some Questions:

1. I have four W's, three of which are in relevant mathematics classes (listed above). I was struggling financially at the time, father was laid off and not able to find steady work. During these periods, often of which were not given with much heads-up, I had to take on 40 - 60 hours of work and could just not keep pace will all of my courses. Figured it probably better to withdraw and retake, rather than take a C. How much weight will this have on my application?

2. Would it be more beneficial to complete a undergraduate thesis OR one year of statistics relevant research with a professor?

 

 

[Wzz]

I am a current student too which means my words might not be helpful. But according to the impression I get from professors, the W's are no problem, as your overall grades shows how strong you are already. There will be a question in application to explain these sort of situations, and you do have a legit reason for these W's. So I guess don't worry?

If I were you, I would do the research. Firstly it will give you a picture of what you like in statistics. Secondly, as I'm doing a thesis for undergraduate now, I feel that at this level a thesis would emphasize more on background research and summarizing what I have already done, which is not as beneficial as doing extra research. Also in speaking of application, I feel that a experience in stats will add more to your profile than a thesis, plus there are probability you finish something within a year and write a paper. Both will be nice, as your math research experience can already endorse your ability. 

About ML and Chicago, maybe adding TTIC to the list?