Thread: Estimate probability of success given k successes in n trials

I'm trying to estimate the probability of success when I'm given data where a certain number of successes (k) occur in a certain number of trials (n).

IOW, instead of trying to figure out the chances of k successes given n trials and a probability (p) (whcih I know how to do). I'm trying to estimate p when given k successes and n trials. An important note is that all potential values for p between 0 and 1 are equally likely.

I can't seem to wrap my head around all of the Bayesian inference stuff I'm reading. I understand exactly what needs to be done involving the integral calculus, I just can't do it. Integration by parts just seems to lead to a never ending process.

Any help?

Originally Posted by Capt Vee

I'm trying to estimate the probability of success when I'm given data where a certain number of successes (k) occur in a certain number of trials (n).

IOW, instead of trying to figure out the chances of k successes given n trials and a probability (p) (whcih I know how to do). I'm trying to estimate p when given k successes and n trials. An important note is that all potential values for p between 0 and 1 are equally likely.

I can't seem to wrap my head around all of the Bayesian inference stuff I'm reading. I understand exactly what needs to be done involving the integral calculus, I just can't do it. Integration by parts just seems to lead to a never ending process.

Any help?

Try Tabular Integration, near the bottom of this page: Integration by parts - Wikipedia, the free encyclopedia

It speeds up things a lot!

Originally Posted by Capt Vee

I'm trying to estimate the probability of success when I'm given data where a certain number of successes (k) occur in a certain number of trials (n).

IOW, instead of trying to figure out the chances of k successes given n trials and a probability (p) (whcih I know how to do). I'm trying to estimate p when given k successes and n trials. An important note is that all potential values for p between 0 and 1 are equally likely.

I can't seem to wrap my head around all of the Bayesian inference stuff I'm reading. I understand exactly what needs to be done involving the integral calculus, I just can't do it. Integration by parts just seems to lead to a never ending process.

Any help?

Hi Capt Vee,

I sounds like you want a Binomial proportion confidence interval. If you simply want a method for computing a confidence interval (rather than a derivation), you can find some methods described here:

Binomial proportion confidence interval - Wikipedia, the free encyclopedia

If this is a Bernoullit trial, try using the Binomial Model.
If it is, let me know. I"ll elaborate.