Reputation
307
Top tag
Next privilege 500 Rep.
Access review queues
Badges
1 11
Newest
 Yearling
Impact
~10k people reached

  • 0 posts edited
  • 0 helpful flags
  • 25 votes cast
Jan
19
awarded  Yearling
Dec
2
revised Expectation due to the cross entropy rate minus the entropy rate
edited tags
Dec
1
asked Expectation due to the cross entropy rate minus the entropy rate
Nov
20
asked Size of a set defined on two multivariate Bernoulli distributions
Oct
30
comment Average Manhattan Distance on a Hypercube
@Did, thanks again. What we now realize is that it is asymptotically $n/2$, after all.... (converging to it very fast).
Oct
30
comment Average Manhattan Distance on a Hypercube
@Eric, please see above my comment to Did.
Oct
30
comment Average Manhattan Distance on a Hypercube
@Did, actually there is an interpretation issue here: what I meant in the OP was the average distance of any two vertices which are different, and ordering does not count (which I think is the natural meaning of average distance of points). Hence, under this interpretation, for $n=1$ the average distance is $1$, not $1/2$; and for $n=2$ it is $4/3$, not $1$.
Oct
30
comment Average Manhattan Distance on a Hypercube
Thanks @Did. That's an elegant answer :)
Oct
30
comment Average Manhattan Distance on a Hypercube
Thanks Eric, but that seems to me to describe the average distance only from a single vertex (0). One has to average over all vertices, without double counting. Also, the normalization $2^n$ seems wrong...
Oct
30
asked Average Manhattan Distance on a Hypercube
Sep
30
revised $Pr(X>Y)<0.5$ iff $E(X)<E(Y)$
added 11 characters in body
Sep
30
asked $Pr(X>Y)<0.5$ iff $E(X)<E(Y)$
Jul
7
asked Inference from $\operatorname{Cov}(X^2,Y)>0$ to $\operatorname{Cov}(X,Y)$
Jul
7
comment Correlation between two sampling RVs
Please see my recent concern, @Did...
Jul
7
comment Absolute value of a sum of non-identically distributed RVs
@Mike Pierce, could this question be removed from "on hold" please?
Jul
7
awarded  Autobiographer
Jul
6
comment Correlation between two sampling RVs
I went back to this problem (actually using the proof for a paper in genetics) and it seems the inequality in your "By convexity $B_i\geqslant1+A_i$" should be the other way around. Please let me know what you think.
Jul
6
comment Absolute value of a sum of non-identically distributed RVs
OK, thanks for your note. I added some thoughts on the problem.
Jul
6
revised Absolute value of a sum of non-identically distributed RVs
edited body
Jul
6
revised Absolute value of a sum of non-identically distributed RVs
added 437 characters in body