Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm interested in various machine learning methods and wonder how can I aim for a career in the field. What are most basic skills to have?

I mean the matters strictly connected to the machine learning - not programming or libraries.

share|improve this question

closed as primarily opinion-based by Martin Thoma, Sean Owen Jan 9 at 9:03

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise.If this question can be reworded to fit the rules in the help center, please edit the question.

I'm sorry I need to vote it close as opinion-based :( Can you edit it and be a bit more specific? – Dawny33 Jan 7 at 12:14
@Dawny33 I mean the matters strictly connected to the machine learning - not programming or libraries. I just don't know how to put it into words. :( – Luke Jan 7 at 12:17
@Luke The problem is that the question "what are the most basic skills" are based on personal opinion. People might disagree on the answer; there is not a single right answer. This kind of question is probably more suitable on – Martin Thoma Jan 7 at 13:26
@moose I wanted to ask for the essential knowledge. Feel free to edit my question. – Luke Jan 7 at 13:31
@Luke I understand what you asked and I've answered to that question. If you think I understood it wrong, you have to clarify. If I got it right, then this is opinion based. (I'm not saying this is a bad question; I only say that it is not suitable for the StackExchange format). – Martin Thoma Jan 7 at 13:36

Machine Learning is a really big field. Depending on what exactly you want to do, there might be huge differences.

Having said that, the following skills are helpful:

  • Programming, especially Python / C++
  • Frameworks like TensorFlow, sklearn, Torch, ...
  • Algorithms like neural networks (especially gradient descent), SVM, decision trees, clustering algorithms, Q-Learning, ...
  • Mathematics (e.g. probability theory / statistics, linear algebra, analysis for calculating gradients

Knowing how to program the GPU directly (CUDA) is probably not always necessary, but certainly a big bonus.

In case you're interested what you learn at KIT (Karlsruhe, Germany) in an introductory machine learning course, see my blog post for exam preparation (I'm sorry, it is mostly in German).

Online Courses

There are lots of online material for machine learning. For example an Coursera ML course

share|improve this answer

I have already answered similar questions here and here

So, that would help you get an idea of how you need to design your own learning path.

The reason why I said that this question is opinion-based is:

Data Science is a huge domain in itself, and attaining complete knowledge of everything in it, is close to impossible.

So, if I am interested in pursuing the concept of Neural Networks alone, and subsequently Deep Learning; then my learning path would be completely different to someone who wants to learn a different concept like time series.

And so does that of someone who wants to learn some other concept. The concepts might be intertwined, but something which I absolutely require in my learning path might not be necessary to someone else's.

So, those above resources are a rough guide to get started.

In addition to that, here is a list of conferences and podcasts of Data Science, which help you with your learning.

And also keep an eye on the tag here.

Additional reading:

Applying machine learning in a real world example A really good neural networks and deep learning book

share|improve this answer

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