You had to use X~B(900,p) and then say that P(a call lasts longer than 30 mins) must be small, so you approximate it as Po(900p), leaving it in terms of p. Then you had that only 5% of days don't have any calls lasting longer than 30 mins (as it said 95% of days do) and so you could say that P(X=0) in a day is equal to 0.05. That means that using your poisson apprximation, you can put the e^minus lambda formula thing equal to 0.05. And then solve for lambda, which you also know equals 900p, and therefore you can find p.
it took me quite a while to work out, literally saw how to do it in about the last minute and a half but luckily got it down.

(edited 10 months ago)

Reply 11

squidgey

7

Original post by bans_x

Who got the last question?

The probabilities were both very close, about 0.55 each or something, but P(red) was slightly higher. For the yellow, you used negative binomial, and had to add all the probabilities P(X=3), P(X=4) ... up to P(X=7), I think those were the right numbers but I can't remember exactly now.

Reply 12

j8ko8

9

Original post by ojha99

I modelled red with geo distribution and yellow with negative binomial. I think red probability was greater

Yh red was greater

Reply 13

j8ko8

9

Geo and negative binomial pretty much the same thing when r=1

Reply 14

j8ko8

9

Original post by squidgey

The probabilities were both very close, about 0.55 each or something, but P(red) was slightly higher. For the yellow, you used negative binomial, and had to add all the probabilities P(X=3), P(X=4) ... up to P(X=7), I think those were the right numbers but I can't remember exactly now.