I have statistical background now I want to learn machine learning. I have experience in R, SPSS, Stata. Necessary resources and tutorials which are effective not only waste my time are needed for my early research career. I have searched several tutorials but I need effective one.

  • $\begingroup$ If you want to get into research work through a serious book like Pattern Recognition and Machine Learning, Machine Learning: A Probabilistic Perspective, or The Elements of Statistical Learning, going through the math and writing as much code as much possible, because researchers must know the theory. You might also want a computer science book, e.g., Sedgewick's Algorithms. After that pick a research problem, learn all about it, then make a small improvement. Your research setting will influence what problem you work on, but you will naturally have your own interests. $\endgroup$ – Emre Jun 15 '17 at 8:33
  • $\begingroup$ I hope you can find suitable resources. However, everyone is different, and will need different source materials and approaches. As Emre points out, what you need to learn depends a lot on where you want to end up. It is possible to spend years diving into single subject matter, or just a week learning how to properly use an ML toolkit. There is no way to really answer this question objectively. $\endgroup$ – Neil Slater Jun 15 '17 at 9:08

I would recommend using Kaggle or a similar website. There are a lot of competition so you probably can find one that motivates you.

One point that is imho often overseen is their very large collection of non-competition datasets. I am rather confident that you will find one that meets a personal interest of yours. You can then start to explore it and discuss results with others.

I really think that machine learning like other programming skills is best learned on practical examples. You will fail, learn from your mistakes only to fall into the next trap and this is very important as these experiences will help you later on!

| improve this answer | |

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