Prospective PhD applicant: should I take real analysis or data mgmt & algor?

bbff1234 发表于 2016-2-9 09:47
IC. Then do you think Intro to Algorithm is a better choice in my case?

No, if you want a PhD in the fields you listed above then go for analysis. What I said before is that grad-level > undergrad-level.

我的电脑能打中文

你一口气提到了很多个领域，probability, stochastic simulation, deep learning, and data mining, dynamical systems
虽然这些领域(除了dynamical system)统计系都有人研究，都需要学很多数学课，但是其所需的知识背景相差很大
比方说dynamical systems需要微分方程（PDE ODE），和stochastic simulation更接近于数学系/计算数学系而不是统计系；Probability需要学大量real analysis，machine learning则不需要

你现在毫无疑问选real analysis。这些方向无论哪一个都是学越多数学越好，
如果选probability那就多学measure based prob, stochastic process..；
machine learning方向学习convex optimization, matrix computation或其他名字有machine learning的课；
dynamical systems学PDE ODE或者研究生级的dynamical systems
stochastic simulation学monte carlo和计算数学相关的课

pigonline 发表于 2016-2-10 10:15
我的电脑能打中文

你一口气提到了很多个领域，probability, stochastic simulation, deep learning, and ...

Thank you so much for your help! Since I'm a sophomore student, I haven't fixed my track yet, although it doesn't hurt to label a few of my personal interests to identify the courses that I need to take.

BTW, is regression analysis very important for any of these topics?

bbff1234 发表于 2016-2-10 11:37
Thank you so much for your help! Since I'm a sophomore student, I haven't fixed my track yet, alth ...

Regression，Probability和Mathematical Statistics是统计最最基本的课，想申统计"系"的PhD必须学啊，否则一点统计背景都没有

real analysis.
probability
linear algebra
regression

and some grad courses would be helpful too.