posted 2 years ago
Hi:
I had an interesting conversation with a friend yesterday.
Between artifacts of NumPy syntax and semantics and the applications, are some general concepts in maths.
I mentioned how difficult I was finding it to envision in detail the re-ordering that occurs when one transposes in more than two dimensions.
In other words, here's my original 3-d, 4-d or higher array, now I do this transposition, where does everything wind up?
Tutorials I looked at were happy enough for us to say what shape results, i.e. the size of each resulting dimension. That part is easy!
My friend claimed he had done data analysis for more than a year and barely ever needed to transpose, I'd been hearing it was extremely common in ML/AI.
How important is transposition of matrices arrays in higher dimensions than two?
I will either invest rather a lot in being able to visualize this / predict what it looks like or very little.
I guess the question should be answered in the scope of your book, but if the answer is "In the scope of this book, not too common, but in these areas -- a lot" I would appreciate that too.
Cheers,
Jesse
RTFJD (the JavaDocs are your friends!) If you haven't read them in a long time, then RRTFJD (they might have changed!)