Confidence Interval - Poisson

Considering X ~ Poi(?). Based on observations x1, ..., x100, iid and with the same distribution as X, we can estimate the confidence interval [0.3;1.8] for ?. Find the confidence interval, with the same coverage probability for P {X<=2}. TIP: express P {X<=2} in terms of ?

  • Extend the deadline for a couple of hours if you can. I may be able to answer this.

  • Max. I can do is +1 hour, is it enough?

Answer

Answers can be viewed only if
  1. The questioner was satisfied and accepted the answer, or
  2. The answer was disputed, but the judge evaluated it as 100% correct.
View the answer

1 Attachment

  • Don't you need to calculate the probability coverage? From what I understood you would need to calculate the probability coverage for the first given interval and then use it to find another interval with P{X<=2} with the same probability coverage. How did you skip not finding it?

  • What do you mean by probability coverage? You know that with certain confidence level lambda in inside the given interval. So with the same confidence level, P[x <=2] will lie in the interval I computed. There doesn't seem to be a need to compute anything else.

  • Sorry, some terms might get lost in translation. Coverage probability I mean the percentage of chance of lambda being inside the confidence interval. So I was thinking you would need to calculate the percentage of chance of lambda being inside [0,3;1,8] and use this same percentage to determine the other confidence interval for P[X>=2]

  • No, that percentage of chance is not given for lambda, and hence you can not find it for p[X<=2]. That percentage of chance is unknown but is the same for both confidence intervals.

  • But how can you be certain that is the same for both?

  • It is because the second interval is directly computed from the first interval, so the probability of both intervals are the same.

  • Makes sense. Why does he mention this coverage probability though? It seems like it wouldn't change your answer with your without it

  • The problem does not ask for the coverage probability. It just wants it to be the same.

The answer is accepted.