Como encontrar

12

Como posso resolver isso? Eu preciso de equações intermediárias. Talvez a resposta seja - t f ( x )tf(x) .

dd t [t xf(x)d x ]

ddt[txf(x)dx]

f ( x )f(x) é a função de densidade de probabilidade.

Ou seja, lim x f ( x ) = 0limxf(x)=0 e lim x F ( x ) = 1limxF(x)=1

fonte: http://www.actuaries.jp/lib/collection/books/H22/H22A.pdf p.40

Tentando equações intermediárias abaixo:

dd t [t xf(x)d x ] = dd t [[xF(x)]t -t F(x)d x ] ? ?

ddt[txf(x)dx]=ddt[[xF(x)]ttF(x)dx]??

dd t a t f(x)d x = - dd t t a f(x)d x = - dd t (F(t)-F(a))=F(t)=f(t)

ddtatf(x)dx=ddttaf(x)dx=ddt(F(t)F(a))=F(t)=f(t)

Hiroaki Machida
fonte
4
Você quer dizer dd t [t xf(x)dx] ? Talvez-tf(t). Ou você quer dizerddt[txf(x) dx]tf(t). dd t [t xf(x)  d x1 - F ( t ) ]? ddt[txf(x) dx1F(t)]
Henry
2
Use o teorema fundamental do cálculo
Henry
2
Consider a primitive GG of xxf(x)xxf(x), then txf(x)dx=G()G(t)txf(x)dx=G()G(t) is easy to derivate.
Stéphane Laurent
2
Please add the self-study tag and read its tag wiki.
Glen_b -Reinstate Monica
2
Se você está estudando para um exame, não é a solução ideal para você. As perguntas de auto-estudo têm como objetivo levar a pessoa que faz a pergunta a resolver o problema sozinha.
Xian

Respostas:

15

Por definição, a derivada ( se existir ) é o limite do quociente de diferença

1h (t + h xf(x)dx-t xf(x)dx)=- 1h t + h t xf(x)dx

1h(t+hxf(x)dxtxf(x)dx)=1ht+htxf(x)dx

como h 0 .h0

Assumindo que f é contínuo dentro de um intervalo [ t , t + h ) para h > 0 suficientemente pequeno , x f também será contínuo ao longo deste intervalo. Então o Teorema do Valor Médio afirma que existe algum h entre 0 e hf[t,t+h)h>0xfh0h for which

(t+h)f(t+h)=1ht+htxf(x)dx.

(t+h)f(t+h)=1ht+htxf(x)dx.

As h0h0, necessarily h0h0, and the continuity of ff near tt then implies the left hand side has a limit equal to tf(t)tf(t).

(It is nice to see that this analysis requires no reasoning about the existence of the original improper integral txf(x)dxtxf(x)dx.)

However, even when a distribution has a density ff, that density does not have to be continuous. At points of discontinuity, the difference quotient will have different left and right limits: the derivative does not exist there.


This is not a matter that can be dismissed as being some arcane mathematical "pathology" that practitioners can ignore. The PDFs of many common and useful distributions have points of discontinuity. For example, the Uniform(a,b)(a,b) distribution has discontinuous PDF at aa and bb; a Gamma(a,b)(a,b) distribution has a discontinuous PDF at 00 when a1a1 (which includes the ubiquitous Exponential distribution and some of the χ2χ2 distributions); and so on. Therefore, it is important not to assert, without careful qualifications, that the answer is merely tf(t)tf(t): that would be a mistake.

whuber
fonte
A very small addendum: There are cases where the integral is differentiable even when f(x)f(x) is not continuous. Let f(x)=0f(x)=0 for x0x0 and f(x)=1f(x)=1 for 0<x<10<x<1 and f(x)=0f(x)=0 for x2x2. Then near 0, F(x)=x2/2F(x)=x2/2 for x0x0 and 0 for x<0x<0, which is perfectly differentiable at x=0x=0.
Alex R.
@Alex Near 0+0+, F(x)=xF(x)=x, not x2/2x2/2. Consider the Fundamental Theorem of Calculus.
whuber
Sorry for the confusion! I define F(x):=xtf(t)dtF(x):=xtf(t)dt.
Alex R.
1
@Alex Your integrand tf(t)tf(t) is continuous near zero, so I fail to see what kind of example you are presenting or what it shows.
whuber
Great derivation (+1) - it might be worth nothing that this result is a case of Leibniz integral rule.
Reinstate Monica
9

Solved...

ddt[txf(x) dx]ddt[txf(x) dx] =ddt[G()G(t)]=ddt[G()G(t)] =ddt[G()]ddt[G(t)]=ddt[G()]ddt[G(t)] =0tf(t)=0tf(t)

Thank you all!!!

Hiroaki Machida
fonte
5
What is G(t)G(t) function? Why the derivative of G()G() is 0?
Vladislavs Dovgalecs