# kelu

 2d comment Infinite discounted sum of weakly dependent Normal random variablesBut this is more a research question, so anything related is interesting to me. So if you have a good idea that requires something slightly different, that's also much appreciated! 2d comment Infinite discounted sum of weakly dependent Normal random variablesWhat I meant is that it is short-range dependent. Let's say it is an ARMA process. 2d asked Infinite discounted sum of weakly dependent Normal random variables Jun 19 comment Expectation on 1/XHow do I see that it doesn't exist? Could you explain that? Thank you! May 12 comment Convex Optimization of quadratic function with inequality constraintsWhat holds if I have strict convexity (you wrote "it" holds, but I'm not sure what you mean)? My question is actually more a research question so it would already help if I could say something about strictly positive definite matrices (excluding the zeros case you mentioned). So let's just assume A is strictly positive definite and symmetric then. Does that help? May 10 comment Convex Optimization of quadratic function with inequality constraints(Take abstract values in the last part of the previous comment). What I don't understand now is, considering that the problem is convex, how comes I have three possibilities? I mean, there can be only one optimum in a convex problem, right? May 10 comment Convex Optimization of quadratic function with inequality constraintsForgot to add that $x_2=1$ in the latter case. By feasibility (because $x_1\geq 1$), 2.) and 3.) can only happen if $sign(b)\neq sign(d)$ or $sign(b)\neq sign(a)$ respectively. Further, it has to hold that $b\geq d$ or $b\geq a$ respectively. May 10 comment Convex Optimization of quadratic function with inequality constraintsI wrote down the problem for the case $n=2$ and tried to solve it using Kuhn-Tucker conditions as you suggested. So suppose $$A=\left(\begin{eqnarray} a & b\\b & d\end{eqnarray}\right).$$ Then I obtain the following three possibilities: $$1.)\quad x_1=x_2=1.$$ $$2.)\quad x_1=1, x_2=-b/d \text{ (if } \lambda_2=0).$$ $$3.)\quad x_1=-b/a\text{ (if }\lambda_1=0).$$ Is this right? May 10 comment Convex Optimization of quadratic function with inequality constraintsHmm, doesn't help, unfortunately that eigenvector isn't 1. But thanks anayway. May 10 comment Convex Optimization of quadratic function with inequality constraintsI think that might be a good tip, thank you (I still don't get it completely, but I have good hopes that I will once I've had a closer look at the Rayleigh quotient). May 9 comment Convex Optimization of quadratic function with inequality constraintsWhy is that? Unfortunately I don't understand that yet. Could you give some more details or do you know a reference? May 9 comment Convex Optimization of quadratic function with inequality constraintsThank you for your answer. I understand that not ALL of the $\lambda_i$ can be zero because then $x*=A^{-1}\lambda=0$. But couldn't SOME of the $\lambda_i$ be zero (and thus SOME of the $x_i$ be something else than 1)? May 8 asked Convex Optimization of quadratic function with inequality constraints Feb 15 comment Gärtner-Ellis for Matrices?Hmm, so maybe this simply doesn't exist? Feb 13 asked Gärtner-Ellis for Matrices? Oct 15 comment When does distribution bootstrap mean converge to distribution sample mean?Yes, sorry, let's assume $\mu$ is finite, forgot that. But does that tell me anything more about my actual question? Oct 14 comment When does distribution bootstrap mean converge to distribution sample mean?What is $\mathbb{P}\left(n^{1/2}\bar{X}_n\leq x\right)$ actually? Is it asymptotically normal? And what can I say about $\mathbb{P}\left(n^{1/2}\bar{X}^*_n\leq x\right)$? Oct 14 comment Limits of series proofs help neededSorry, I meant $b_n$ is not in this interval for $n\geq n_\epsilon$, of course. Oct 14 answered Limits of series proofs help needed Oct 14 asked When does distribution bootstrap mean converge to distribution sample mean?