I am interested in a few areas with biostatistics, and was reading the course catalog at a university that offers machine learning. I am taking topology now, and i think machine learning uses topology (not sure), but is machine learning very useful with cancer analytics? if so, is what fields of math does it require to study machine learning? does it only require probability theory?

For me right now I am going into graduate studies in biostats, and am interested in learning Markov modeling, regression modeling, and linear algebra. I think linear algebra is very very useful for doing multivariable analytics and large data sets...etc.

my concern is that I can not master everything! It would be too difficult to master linear algebra, machine learning.... statistical regression etc...

So I just am inquiring about input regarding machine learning, and how to prepare for it, if and only if, it is highly regarded ... thank you



closed as too broad by Nick Cox, Andy, Scortchi - Reinstate Monica, gung - Reinstate Monica, whuber Oct 24 '14 at 15:11

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    $\begingroup$ Machine learning has [almost] nothing to do with topology! To be comfortable studying machine learning you would probably want to be familiar with basic first-year undergraduate mathematics, including linear algebra, analysis (aka calculus) and some probability theory, but not much else is needed. $\endgroup$ – amoeba says Reinstate Monica Oct 24 '14 at 13:23

In my opinion, when you are still learning, you should allow yourself to be exposed to many topics but focus more detailed effort (i.e. course selection, projects) on something that really interests you. Don't be overly influenced by the flavour-of-the-month. If you really master and dive deep into one specialized branch of analysis, you'll find learning the next one much easier as you'll be able to anchor yourself in that initial area you found personally interesting. The same principles will come up over and over, even if the terminology might be different from one branch to another.

In practice, I have found the data dictates the approach. No, you can't master everything, but if you know the existence of different approaches, and have that initial strong foundation in something, you'll be well equipped to "master" other approaches in the future and choose them appropriately. Graduate school will generally proceed such that you'll come out the other side with this ability.

"Cancer analytics", and "machine learning" for that matter, are pretty broad terms but, yes, different variations of those two things come together in many different ways and is an extremely active area with more than enough room for keen initiates.

  • $\begingroup$ The reason i ask is not for assurance. but for a clear path. for instance, I once had lunch with a physics PhD graduate and his field of study had no funding from the state department. he was working on material science, and his specialty was being funded $0 dollars. What does this mean? I am interested of course in creative ideas, but if the ideas have no interested or value to others, I think it best to select a different topic to explore. i just wish to avoid a barren wasteland :) $\endgroup$ – sophie-germain Oct 24 '14 at 8:14
  • $\begingroup$ I can think of several techniques that came and went in cancer genomics, but it certainly didn't leave any colleagues that used them in funding trouble when they fell out of favour. They just moved on to better techniques (or created new ones themselves). I can think of many active projects in cancer research for all of the topics you've mentioned. You'd have to define a topic at a much finer level before someone could bring up lack of interest or funding. Good luck! $\endgroup$ – scottzed Oct 24 '14 at 13:47

There are already publications that try to apply machine learning techniques in cancer research. It is usually for identifying specific types of cancer that respond to specific treatments. Check pubmed for oncology and machine learning. Also, cancer research is typically well-funded.

For proper understanding, you'll have to study first linear algebra, then statistical regression and then machine learning.

As a personal note, I started my studies with a specific goal like you. I ended up with something completely different and feel great with it.


Not the answer you're looking for? Browse other questions tagged or ask your own question.