288 reputation
211
bio website google.com
location North Pole
age 15
visits member for 2 years, 8 months
seen Jul 16 at 15:31

Prefers to stay anonymous and eat mousse au chocolat.


Jun
19
awarded  Nice Answer
Jun
6
comment Pearson correlation and metric properties
Interesting idea. z-standardization is a linear transformation in the transposed space. If metric properties survive transposition, this would pretty much yield a proof.
Feb
26
comment Precision and performance of Euclidean distance
I don't get any precision problems on this example.
Feb
21
awarded  Notable Question
Feb
5
awarded  Commentator
Feb
5
comment Is it faster to count to the infinite going one by one or two by two?
To make it more explicit: when everybody moves from room $i$, to room $2i$ - who will be in a room with an odd number afterwards? The room must have become empty! And when you are told to move to room $2i$, that room must also be empty - because whoever was in there just moved to room $4i$. Everbody moves to a room that just became vacant.
Feb
5
comment Is it faster to count to the infinite going one by one or two by two?
Try to prove that the room is occupied. Because whoever was in there, supposedly moved to room $\pi(\pi(j))$, didn't he? Therefore, the destination room became empty.
Feb
5
revised Is it faster to count to the infinite going one by one or two by two?
added 994 characters in body
Feb
4
revised Is it faster to count to the infinite going one by one or two by two?
added 73 characters in body
Feb
4
awarded  Yearling
Feb
3
awarded  Teacher
Feb
3
answered Is it faster to count to the infinite going one by one or two by two?
Feb
1
comment Estimating the geometric shape of a point cloud without using the vertex information
Define "geometric shape". Have you looked at alpha shapes? I don't think clustering is what you are looking for. Why would you lose vertex information the first place?
Jul
17
awarded  Popular Question
Jun
2
comment Cosine similarity / distance and triangle equation
Note that in C I asked about it being pseudo metric: en.wikipedia.org/wiki/Pseudometric_space which does allow non-identical vectors to have distance 0.
Feb
24
awarded  Popular Question
Feb
6
asked Pearson correlation and metric properties
Jan
5
comment rewriting to avoid catastrophic cancellation
No. The answer by @penartur is correct. If x^2 approximately is the same as y^2, and large, you can lose a lot of signficant digits.
Oct
8
awarded  Tumbleweed
Jun
13
comment Fast approximate construction of orthogonal system
I'm fine with any other similar method, too, that gives me "direction" vectors and associated variances (which should be computable in O(n^2) anyway). I need reasonable candidates, but faster than the common exact methods for matrix inversion or eigenvector decomposition which all are O(n^3).