after fundamentals what is the next course for a physics student?

21 views (last 30 days)
Hi, I am on chapter 12 of 15 total chapters in Fundamentals of MatLab. I feel I am only skimming the surface of how matlab skills need to be developed in doing science models etc. After I finish this 15th chapter, is there a way to continue my training? What course is recommended? Thank you.

Accepted Answer

Cedric
Cedric on 3 Jun 2014
Edited: Cedric on 3 Jun 2014
I don't have much time but I'll bring my 2 cents too, and develop two points.
A. You have to make the distinction between the science that your are trying to tackle/study, and the technical aspect of programming with MATLAB. Most important, you have to realize that, once you'll be comfortable enough with programming techniques, your issues will lie at the interface between the two: you will have to determine whether your numerical computations really relate to what you are studying, how to evaluate the error, etc. In other words, it is generally easy to obtain numbers, but extremely difficult to evaluate their relevance or even the scope of their validity.
B. Yet, in research (= science made pragmatic in a capitalist context?), you'll have to be fast when you need to test ideas/approaches. If it takes an hour to investigate a computational approach which includes a bit of data handling, a bunch of computations, and the creation of relevant outputs, you can evaluate 50 approaches over a 2 days period and pick the best suited for your purpose, without messing up the deadlines of your funding project(s). If it takes a week or a month, because you don't know how to handle data files/sources and have to switch between MATLAB and Excel repeatedly because you cannot work without Excel's VLOOKUP, you are likely to stick to the first method that comes to your mind and try to sell whatever comes out of it, because there is no time and money for more than that. Therefore, as far as I am concerned, it is important to grow your technical skills until you are able to perform fast prototyping for testing ideas and approaches. This is also how you can investigate and develop new research directions for which you have no funding (yet!).

More Answers (5)

Image Analyst
Image Analyst on 2 Jun 2014
Yes, I recommend you go into image analysis. It will be a fun and rewarding field for you with a lot of variety of projects you can work on. Or you can work in one field forever (like medical, fingerprints, remote sensing, etc.) if you want, it's up to you. And you can program in MATLAB almost all day long. Of course my opinion is very biased and I'm sure there are other fields you may find more interesting. You can look over Steve's book on MATLAB here: http://imageprocessingplace.com/DIPUM-2E/dipum2e_main_page.htm. His page is at http://blogs.mathworks.com/steve/
  1 Comment
Chad Greene
Chad Greene on 2 Jun 2014
Excellent point. In every field I've worked in, even underwater acoustics, I've encountered problems where I've wished I had a better understanding of image analysis tools. Image processing expertise will serve you well in any field.

Sign in to comment.


Star Strider
Star Strider on 2 Jun 2014
I can’t really argue with Chad Greene’s or Image Analyst’s suggestions. You can combine them by traveling and taking photographs, then working with them when you return.
For a next course though, I would suggest statistics and experiment design, including probability theory if you haven’t already had it. MATLAB’s Statistics Toolbox is powerful, but it’s only as powerful as your knowledge and understanding of the underlying principles. Competency in statistics will allow you to design efficient experiments, and it will provide you with the knowledge to differentiate good and bad statistical design and analysis in the papers you read, and with that determine whether the inferences and conclusions in them are valid. You might well see things that the reviewers and editors missed.
  2 Comments
Image Analyst
Image Analyst on 2 Jun 2014
Yes, statistics would be a great addition. I wish I knew more about stats. Luckily (or unluckily) I don't know as much as I want to because we have an army of statisticians (well okay, maybe just a few dozen) at our company at my beck and call who can do the complicated things way better than I ever could. They're always offering internal classes and I take them whenever I can, not so I can do the stats myself (wow, that would be risky and dangerous!) but so I can talk somewhat intelligently with them in the same jargon when they're working on my projects. I've taken courses on cluster analysis, design of experiments, gauge R&R, partial least squares, etc. But they don't know about the myriad of optimization methods. If you ask them when should you use or note use adaboost, treebagger, bag of words, support vector machines, particle swarm optimization, ant colony optimization, RANSAC, etc. you might just get a "huh?"
Chad Greene
Chad Greene on 2 Jun 2014
A good point by Star Strider. Numbers are meaningless without statistics. I keep hearing how good this book is. Very readable.

Sign in to comment.


Chad Greene
Chad Greene on 2 Jun 2014
My personal opinion is if you've gone through that whole book you've got more than enough foundation to start tackling any problem, and this forum is here for you if you ever get stuck. For me, the best learning happens when I have real-world problems to solve. So what should be next? If I were you I'd close my books, turn off my computer, spend the summer traveling, and learn more about Matlab when your research warrants it. But everyone has their own style, so do what suits you.
  1 Comment
Christopher Raabe
Christopher Raabe on 2 Jun 2014
thank you for this thought; i will finish the next three chapters as soon as I am able and I certainly am optimistic about future learning. much thanks, cwr

Sign in to comment.


Sean de Wolski
Sean de Wolski on 2 Jun 2014
How about getting a book on finance, making millions of dollars, and then doing whatever you find to be the most interesting?
I feel I should put the disclaimer: These are my opinions alone, they do not reflect those of my employer and do not come with any warranties or provisions indicating that you would actually make any money .
  2 Comments
Image Analyst
Image Analyst on 2 Jun 2014
Somewhere I heard that more engineers are going to work in the financial industry than the semiconductor and telecommunications industries combined - something like 25% at MIT for example.

Sign in to comment.


Chad Greene
Chad Greene on 3 Jun 2014
Better than books if you're trying to get comfortable with Matlab and/or refine your skills: Cody. I have not participated, but the premise seems perfect for learning tricks and increasing your efficiency.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!