d2jsp
Log InRegister
d2jsp Forums > Off-Topic > General Chat > Homework Help > Stats Problem
Add Reply New Topic New Poll
Member
Posts: 4,076
Joined: Aug 20 2007
Gold: 0.00
Oct 12 2013 11:01pm
Hiya,

So I don't exactly want a direct answer to this problem, but if someone could just explain what exactly is going on or how to get started, or where to look to get more in formation, that would be greatly appreciated. I don't feel my professor has done a great job instructing, and I haven't read the book as much as I probably should have due to having other obligations.

Anyway, here's the problem:



Thanks in advance!
Member
Posts: 4,076
Joined: Aug 20 2007
Gold: 0.00
Oct 14 2013 12:59am
Bumping this. Anyone have any advice? :(
Member
Posts: 16,662
Joined: Nov 24 2007
Gold: 15,245.00
Trader: Trusted
Oct 14 2013 11:36am
Just to start you :

1- No idea what you're precisely asked.


2- T = max{X1...Xn}
P ( 0<T<A) = ( A / thêta ) ^ n
since we need all X1 ... Xn to be smaller than thêta for their max to be smaller than thêta.

Hence : P ( 0<T<A ) = Integral ( t=0 to A ; f(t).dt )
with f(t) = n*t^(n-1) / (thêta^n)

E(T) = Integral ( t=0 to thêta ; t*f(t).dt ) = ... easy calculations ... = n*thêta / (n+1)

I suppose your constant c should be c = (n+1) / (2n)
such that E(U) = c*E(T) = thêta / 2

V(U) = c² * V(T) = ... not so hard calculations ... = thêta² * n / [ (n+1)² * (n+2) ]

but you will need help from another one with Cramer-Rao bound :(
Member
Posts: 4,076
Joined: Aug 20 2007
Gold: 0.00
Oct 16 2013 01:47pm
Thank you for the help! :D

Still trying to figure out the first one though. Gonna spend the next two days carefully reviewing the text.
Go Back To Homework Help Topic List
Add Reply New Topic New Poll