Reputation
20,101
Next tag badge:
97/100 score
56/20 answers
Badges
1 13 41
Newest
 Tag Editor
Impact
~152k people reached

19h
accepted Improving the performance of eigs for a large spd Problem
19h
comment Linear Transformation on $\mathbb{R}^6$
@Diya You're right, option (D) is true (since $\dim \ker T > 0$). I don't expect the answerer to be around anymore though looking at his rep and last seen...
1d
answered Improving the performance of eigs for a large spd Problem
1d
comment Improving the performance of eigs for a large spd Problem
@uranix Finally came up with a solution. Your idea gave me the hint :) I'll write it up as an answer.
2d
comment Improving the performance of eigs for a large spd Problem
@uranix The Problem is I'll have to use the iterative approach because my memory doesn't fit a full matrix of that size. I'm not faniliar enough with the Lanczos algorithm to know how to accelerate it.
2d
comment Improving the performance of eigs for a large spd Problem
@uranix I dont have some Terabytes of RAM ^^
2d
comment Improving the performance of eigs for a large spd Problem
@uranix Thanks, I'll try that. eig didn't allow sparse input by the way.
2d
comment Improving the performance of eigs for a large spd Problem
@horchler Yes, even 32GB of RAM can't hold a full matrix of that size ^^ I'll try eig as soon as I get home, but I thought it needs full matrices.
Aug
24
revised Improving the performance of eigs for a large spd Problem
added 32 characters in body
Aug
24
awarded  Tag Editor
Aug
24
wiki created sparse-matrices description
Aug
24
wiki created sparse-matrices excerpt
Aug
24
asked Improving the performance of eigs for a large spd Problem
Aug
22
comment sum of fractional functions optimization problem
fractional-calculus is a completely different matter, hence I removed the tag. For the problem: What else do you know about $X$? Is the objective function self-concordant on $X$? If so, you can apply standard convex optimization theory.
Aug
22
revised sum of fractional functions optimization problem
edited tags
Aug
22
comment Do convergence a.e. + limit function being in $L^p$ imply $L^p$ convergence?
@user254665 That's the $p=2$ case of Daniel Fischers' general example.
Aug
22
comment Do convergence a.e. + limit function being in $L^p$ imply $L^p$ convergence?
@DanielFischer Nice counter-example for the second question. Unfortunately the OP deleted the comment. Could you undelete it so your comment remains clear?
Aug
22
answered Solving a system of polynomial equations
Aug
22
comment what does “matrix valued function is continuous differentiable” mean?
@Sean A matrix valued function would have codomain $\mathbb R^{n\times m} \simeq \mathbb R^{n\cdot m}$. Your codomain just consists of pairs of vectors, one from $\mathbb R^q$ and one from $\mathbb R^d$.
Aug
17
comment Determining a confidence interval
You need to specify the assumed distributions for your variables. They cannot be normally distributed (what the test you used assumes), since they must clearly be discrete and supported on $[0,N]$, not continuously on $\mathbb R$.