4 ms·
ISLR is a good one—it’s focused on practical theory, but there’s still some unavoidable math. Code exercises are simple (and ugly—stats professor code) but give
by jointpdf 8y ago
ISLR is a good one—it’s focused on practical theory, but there’s still some unavoidable math. Code exercises are simple (and ugly—stats professor code) but gives you the basics to build on.
Book/code: http://www-bcf.usc.edu/~gareth/ISL/ http://www-bcf.usc.edu/~gareth/ISL/
Lectures: https://www.r-bloggers.com/in-depth-introduction-to-machine-learning-in-15-hours-of-expert-videos/amp/ https://www.r-bloggers.com/in-depth-introduction-to-machine-...
- jointpdf 8y agoAlso: this was recommended on HN recently and it’s the best intro to matrix algebra / basic math of ML that I’ve seen. Lots of very practical and relevant exercises that relate to real data science tasks like text and image analysis. Code examples are in Julia. http://vmls-book.stanford.edu/ http://vmls-book.stanford.edu/ And one more gold standard for more practical data science (with R) stuff. It introduces a more modern take on R (chaining functions together with the %>% pipe operator, which makes everything clean and terse). https://r4ds.had.co.nz/ https://r4ds.had.co.nz/