码工Interview的数学问题，what's the non positive definite？傻了

所有跟帖： 

• Quadratic forms or symmetric matrices? -BeLe- ♀ (0 bytes) () 03/06/2015 postreply 16:27:09

• 查了一下，和matrix有关，positive definite比较容易理解，non positive definite不容易理 -flagsix- ♂ (6 bytes) () 03/06/2015 postreply 17:01:40

• 比如 0-矩阵 -weston- ♀ (0 bytes) () 03/06/2015 postreply 17:03:25

• It is also non-negative definite. -career- ♂ (41 bytes) () 03/06/2015 postreply 20:27:07

• You don't get it. No surprise here. -weston- ♀ (0 bytes) () 03/07/2015 postreply 06:14:59

• I got your not-so-good math skill. :) -career- ♂ (0 bytes) () 03/07/2015 postreply 06:59:58

• In term of symmetric matrices -BeLe- ♀ (231 bytes) () 03/06/2015 postreply 18:00:39

• 学习了, 多谢! -flagsix- ♂ (0 bytes) () 03/06/2015 postreply 18:09:07

• The opposite of positve is negative. Non-positve symmetric -weston- ♀ (160 bytes) () 03/06/2015 postreply 18:15:59

• It's only wrong b/c you missed the context. -weston- ♀ (0 bytes) () 03/06/2015 postreply 21:23:03

• Your old defense trick again! :) -career- ♂ (0 bytes) () 03/06/2015 postreply 21:27:41

• You are as right here as your "sinc(x) is NND". What a joke. -weston- ♀ (0 bytes) () 03/07/2015 postreply 06:10:39

• Yes, sinc(x) is non-negative definite, because its Fourier Tran -career- ♂ (212 bytes) () 03/07/2015 postreply 06:42:57

• Bochner calls it "positive definite", not "NND". Get a better te -weston- ♀ (0 bytes) () 03/07/2015 postreply 06:56:21

• FT(sinc(x)) becomes zero at lots of points, doesn't it? -career- ♂ (53 bytes) () 03/07/2015 postreply 07:06:36

• Sigh. You don't understand the proof of the Bochner Theorem. Go -weston- ♀ (0 bytes) () 03/07/2015 postreply 09:05:12

• You don't need more "proof"--it's by the definition! :) -career- ♂ (0 bytes) () 03/07/2015 postreply 09:15:36

• Surprised anyway would hire someone like you who doesn't know -weston- ♀ (78 bytes) () 03/07/2015 postreply 09:44:02

• You forgot any integral over a zero-valued set yields zero. :) -career- ♂ (62 bytes) () 03/07/2015 postreply 11:18:28

• Your step function is not zero over an open interval. Check Four -weston- ♀ (0 bytes) () 03/07/2015 postreply 11:21:46

• It's a rectangular function--not a "step"! -career- ♂ (67 bytes) () 03/07/2015 postreply 11:33:18

• Wikipedia "rectangular function, the next simplest step function -weston- ♀ (0 bytes) () 03/07/2015 postreply 11:40:53

• hoho, you even need Wiki for that! :) -career- ♂ (168 bytes) () 03/07/2015 postreply 11:46:47

• For your conenience only, since you had idea step functions cove -weston- ♀ (0 bytes) () 03/07/2015 postreply 11:57:00

• I have the rectangular. it is you who "mis"-"step":))) -career- ♂ (567 bytes) () 03/07/2015 postreply 11:59:34

• 你这是 "白马非马" 的现代版 -weston- ♀ (0 bytes) () 03/07/2015 postreply 12:03:16

• Stay in math,pls. :) -career- ♂ (0 bytes) () 03/07/2015 postreply 12:19:30

• :( -career- ♂ (38 bytes) () 03/06/2015 postreply 20:38:52

• was already said -weston- ♀ (0 bytes) () 03/06/2015 postreply 21:24:44

• This? :) -career- ♂ (677 bytes) () 03/06/2015 postreply 21:47:52

• It also can be of a function :) -career- ♂ (213 bytes) () 03/06/2015 postreply 19:53:55

• " sinc(x) is non-negative definite."--What? -weston- ♀ (0 bytes) () 03/06/2015 postreply 21:24:01

• FT(sinc(x)) is non-negative. :) -career- ♂ (0 bytes) () 03/06/2015 postreply 21:30:52

• Well, it's not a typo, since you wrote it twice. Your 1st msg wa -weston- ♀ (316 bytes) () 03/07/2015 postreply 06:09:11

• Not a typo; nor any mistake. You are so lost. :) -career- ♂ (685 bytes) () 03/07/2015 postreply 06:49:54

• see my Bochner comment above. Your behavior continues to prove m -weston- ♀ (0 bytes) () 03/07/2015 postreply 06:57:40

• :))) -career- ♂ (523 bytes) () 03/07/2015 postreply 07:14:41

• Instead of spitting nonsense here, you should go to understand B -weston- ♀ (39 bytes) () 03/07/2015 postreply 09:08:04

• More likely, you misquoted that guy. :) -career- ♂ (64 bytes) () 03/07/2015 postreply 09:18:39

• Your memory of constructing positive definite funciton is failin -weston- ♀ (101 bytes) () 03/07/2015 postreply 09:41:18

• You forgot any integral over a zero-valued set yields zero. :) -career- ♂ (62 bytes) () 03/07/2015 postreply 11:19:23

• No. But the integral takes place over the entire space. Silly -weston- ♀ (0 bytes) () 03/07/2015 postreply 11:22:38

• Then look at your insisting on including the zero-valued subspac -career- ♂ (125 bytes) () 03/07/2015 postreply 11:29:43

• You learned Fourier by skipping all non-zero values of a functio -weston- ♀ (0 bytes) () 03/07/2015 postreply 11:42:21

• Show me where I did that? :) -career- ♂ (0 bytes) () 03/07/2015 postreply 11:48:47

• Out of (-.5, 0.5), rect's Lebesgue measure is zero:) -career- ♂ (200 bytes) () 03/07/2015 postreply 12:16:05

• Repeat one more time: Go and study Bochner Theorem (actually -weston- ♀ (35 bytes) () 03/07/2015 postreply 12:26:42

• Bochner/ Lebesgue won't CYA :) -career- ♂ (813 bytes) () 03/07/2015 postreply 12:40:48

• Gave you the math definition already. Your comment adds nothing -weston- ♀ (0 bytes) () 03/07/2015 postreply 13:07:01

• Got your "Say it the 3rd time, rec funcition is a step function. -career- ♂ (80 bytes) () 03/07/2015 postreply 13:10:25

• Too bad, you can't read math -weston- ♀ (0 bytes) () 03/07/2015 postreply 14:22:16

• Sounds like an interview at MathWorks :) -career- ♂ (0 bytes) () 03/06/2015 postreply 21:22:35