Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I think PCA is better understood in theorem-proof style of mathematics rather than algorithmically. For example, PCA can be understood as an optimization problem. Whereas CART and k-NN and the likes are best understood algorithmically. I know k-NN was introduced as an algorithm and only later figured out in terms of probability, etc..


Algorithmically, it's just finding an eigensystem...


Yes, but what's interesting about PCA is mathematical, not algorithmic.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: