# S2 Poisson - any clever ways to work backwards

Watch
Announcements

Page 1 of 1

Go to first unread

Skip to page:

I had a question where I had to work backwards and found an approximate answer but it was off by quite a bit. So I was wondering if there is a more efficient way to find the value I need.

The question is:

1 person in 80 has blood type P.

In first part of question a sample of 150 people is taken, hence mean is 150 * 1/80 = 1.875 and so you can use X~Poisson(1.875) to calculate probabilities.

But the last part of the question asks:

A hospital urgently requires blood type P. How large a random sample of donors must be taken in order that the probability of finding at least one donor of type P should be 0.99 or more?

This is how I attempted it.

We need P(X >= 1) >= 0.99

and Poisson(x/80, 0) <= 0.01

I just did it trial and error entering mean as 1,2,3,4,5 and got to a probability that matched.

I got to Poisson(5, 0) = 0.00674 for answered that 5 * 80 = 400+ should be sampled.

But this was not very accurate and optimal value is 369.

ie Poisson(369/80, 0) is perfect value.

Is there a more efficient way to find the value than just trying lots of values???

The question is:

1 person in 80 has blood type P.

In first part of question a sample of 150 people is taken, hence mean is 150 * 1/80 = 1.875 and so you can use X~Poisson(1.875) to calculate probabilities.

But the last part of the question asks:

A hospital urgently requires blood type P. How large a random sample of donors must be taken in order that the probability of finding at least one donor of type P should be 0.99 or more?

This is how I attempted it.

We need P(X >= 1) >= 0.99

and Poisson(x/80, 0) <= 0.01

I just did it trial and error entering mean as 1,2,3,4,5 and got to a probability that matched.

I got to Poisson(5, 0) = 0.00674 for answered that 5 * 80 = 400+ should be sampled.

But this was not very accurate and optimal value is 369.

ie Poisson(369/80, 0) is perfect value.

Is there a more efficient way to find the value than just trying lots of values???

0

reply

Report

#2

(Original post by

I had a question where I had to work backwards and found an approximate answer but it was off by quite a bit. So I was wondering if there is a more efficient way to find the value I need.

The question is:

1 person in 80 has blood type P.

In first part of question a sample of 150 people is taken, hence mean is 150 * 1/80 = 1.875 and so you can use X~Poisson(1.875) to calculate probabilities.

But the last part of the question asks:

A hospital urgently requires blood type P. How large a random sample of donors must be taken in order that the probability of finding at least one donor of type P should be 0.99 or more?

This is how I attempted it.

We need P(X >= 1) >= 0.99

and Poisson(x/80, 0) <= 0.01

I just did it trial and error entering mean as 1,2,3,4,5 and got to a probability that matched.

I got to Poisson(5, 0) = 0.00674 for answered that 5 * 80 = 400+ should be sampled.

But this was not very accurate and optimal value is 369.

ie Poisson(369/80, 0) is perfect value.

Is there a more efficient way to find the value than just trying lots of values???

**acomber**)I had a question where I had to work backwards and found an approximate answer but it was off by quite a bit. So I was wondering if there is a more efficient way to find the value I need.

The question is:

1 person in 80 has blood type P.

In first part of question a sample of 150 people is taken, hence mean is 150 * 1/80 = 1.875 and so you can use X~Poisson(1.875) to calculate probabilities.

But the last part of the question asks:

A hospital urgently requires blood type P. How large a random sample of donors must be taken in order that the probability of finding at least one donor of type P should be 0.99 or more?

This is how I attempted it.

We need P(X >= 1) >= 0.99

and Poisson(x/80, 0) <= 0.01

I just did it trial and error entering mean as 1,2,3,4,5 and got to a probability that matched.

I got to Poisson(5, 0) = 0.00674 for answered that 5 * 80 = 400+ should be sampled.

But this was not very accurate and optimal value is 369.

ie Poisson(369/80, 0) is perfect value.

Is there a more efficient way to find the value than just trying lots of values???

This is .

This can be solved using logs.

0

reply

(Original post by

You can solve this using the binomial distribution. You need the probability that no type P donor is found to be less that 0.01.

This is .

This can be solve using logs.

**BuryMathsTutor**)You can solve this using the binomial distribution. You need the probability that no type P donor is found to be less that 0.01.

This is .

This can be solve using logs.

OK, I think I have got it. Because r = 0 you can do as:

nC0 (79/80)^n-0 (1/80)^0

nC0 = 1

so simplifies to 79/80^n as you stated.

0

reply

X

Page 1 of 1

Go to first unread

Skip to page:

### Quick Reply

Back

to top

to top