find the n that maximizes P(A)
$$P(A) = {{\left( {\matrix{ {10} \cr 4 \cr } } \right)\left( {\matrix{ {n - 10} \cr {16} \cr } } \right)} \over {\left( {\matrix{ n \cr {20} \cr } } \right)}}$$
maximum P(A) when (n,20 ) minimum right ? like when its 1 for example ? how ever that wont work since n-10 must be >16
how are you going by solving this?
wait thinking
@ganeshie8
This is what we're talking about right?\[\Large \left(\begin{matrix}p \\ k\end{matrix}\right)=\frac{p!}{(p-k)! \ k!}\]
yep
wanna expand ??
there is a rule how to get the largest(n,k) value by controling k to chose k in the middle to get the middle triangle number which also the largest
but that would be like trial /error if we did it
So these are all little pascal's triangles. That means the middle one will be the largest since we know the outside is all 1's right? Since we can only have a factorial of 0 or greater, that puts some limits on what n can be. Since we have n-16 above a 10 then we absolutely have to have at least n>=26
sowe would have least n of 22 ?
Expanding it out makes it a little clearer. \[\Large \frac{10!}{(10-4)!*4!} \frac{(n-10)!}{(n-26)!*16!} \frac{(n-20)!*20!}{n!}\]
\[\frac{ f(n+1) }{ f(n) }>1; \frac{ f(n+2) }{ f(n+1) }<1\] \[\frac{ f(n+1) }{ f(n) }=\frac{ \left(\begin{matrix}n-19 \\ 16\end{matrix}\right) }\left(\begin{matrix}n \\ 20\end{matrix}\right){ ?\left(\begin{matrix}n+1 \\ 20\end{matrix}\left(\begin{matrix}n-10 \\ 16\end{matrix}\right)\right) }\]
fkkkkkkk
i feel it should be 21 or 22 mmm
$${{f(n + 1)} \over {f(n)}} = {{\left( {\matrix{ {n - 19} \cr {16} \cr } } \right)\left( {\matrix{ n \cr {20} \cr } } \right)} \over {\left( {\matrix{ {n + 1} \cr {20} \cr } } \right)\left( {\matrix{ {n - 10} \cr {16} \cr } } \right)}} = {{{\textstyle{{(n - 9)!} \over {16!(n - 25)!}}}{\textstyle{{n!} \over {(n - 20)!20!}}}} \over {{\textstyle{{(n - 10)!} \over {16!(n - 26)!}}}{\textstyle{{(n + 1)!} \over {(n - 19)!20!}}}}}$$
what ur trying to do nin ?
committing a mathematical abomination? laughing out loud IDK
a few cancelations can be done and rewriting
16! 20! cancel out
ok ill do trial error /last option lol ok so we need maximum (n-10,16) and minimum (n,20) n>=26 for max 16*2=32 so n=42 mmmm
\[=\frac{ (n-9)!n! (n-26)!(n-19)! }{ (n-10)!(n+1)!(n-25)!(n-20)! }\]
It's not that complicated really if we just think of these two parts separately with the n's in them. Since just simply looking at it shows us n has a lower possible bound of 26, let's just do a change of variable to help us look at what we really need: n=26+k \[\Large \frac{10!}{6!*4!} \frac{(k+13)!}{(k)!*16!} \frac{(k+6)!*20!}{(26+k)!}\] This looks more manageable. Let's make it more manageable and scoot all the stuff without k's to the left and the stuff with k's to the right. \[\Large \frac{10!*20!}{6!*4!*16!}*\frac{(k+6)!*(k+13)!}{k!(26+k)!}\]Alright so what's going to maximize the right part? The rest are constants so it really doesn't matter. It looks like we can simplify some of these down to get: \[\Large \frac{(k+1)(...)(k+6)}{(k+14)(...)(k+26)}\] Yeah now take the derivative, set it equal to zero and uhhhhh... lul idk
laughing out loud kai
so maximize function ?@ganeshie8
ok ok 10 and 4 there for a reason what course ur taking @nincompoop is it calcules or descrete math ?
\[= \frac{ (n-9) (n+9)}{ (n+1)(n-25) }\rightarrow (n-9)(n-19)>(n+1)(n-25)\]
\[49>n\]
so I was dreaming about not knowing what I am doing in math, and I think this was it
Wait! I just think I figured it out. So remember if we set k=0 then that's the lowest possible number our thing can have, so that will give us: \[\Large \frac{6!*13!}{26!}\] Now just try every value from here on up haha. I'm going to bed, good luck everyone, may the force be with you etc... lol.
wait can we make usfull of this identity? (n,k) (k,r)=(n,r)(n-r, k-r)
go ahead @ikram002p
mmmm see this :-\ http://www.wolframalpha.com/input/?i=maimize+%28 \binom{10}{4}*\binom{n-10}{16}%29%2F\binom{n}{20}
also table from 26 to 42 gives rational numbers mmm
as ic in the graph n btw 48 to 50
according to that result, the max should be 50, which looks like it corresponds with the graph @ikram002p showed
50 to 48 but not sure mmmm lets see value of it when n>100
49
48, but n+1, so 49
that was only trial error :-\ just tell what course u have ?
probability and statistics
http://math.arizona.edu/~jwatkins/o-mle.pdf shows a method of maximum likelihood estimation (example 5)
supposedly, an intro
shiz that's with an R software, finally! laughing out loud
I suppose you can just use R yeah to find the max. They do go through a proof though and they show \[ \hat{n}=\lfloor \frac{10(20)}{4}\rfloor\]
I don't know how to use it yet, but I am learning this stuff and subject for the first time
nice !
Hm lemme see if I can get a code that will work. Hm that is a difficult question for a first time o.o
I tried this in R: > optimize(fn, interval=c(40, 60), maximum=TRUE) $maximum [1] 49.49498 $objective [1] 0.2801753 I'm not sure which algorithm R uses though for this function.
@nincompoop
oh and of course, fn<-(choose(10,4)*choose(x-10,16))/(choose(x,20))
Join our real-time social learning platform and learn together with your friends!