[albus]

This exam is tomorrow.
Is everyone ready.

[Benniboi1]

Not in the slightest. Tonight should be fun.

[albus]

Has everyone learnt the sampling techniques.

[albus]

This is my last exam

[albus]

(Original post by Amanda.P)
My last exam too, i'm not ready at all though. There are no resources to learn this except the book and that bores me

Try the Integralmaths website, its pretty good for revision.

[james20207]

Hey everyone, I literally started learning this a week ago! Does anybody fancy explaining to me how you use the chi-squared to test the goodness of fit and how you work out the expected frequencies and that? That's the only part I don't understand!

[lubamira]

I've left doing work for this exam way too late! (I had C3 and FP2 today so I needed to concentrate on those) However, I was just wondering if anyone knew anything about interpolation? I've only just heard about it today!

[albus]

(Original post by james20207)
Hey everyone, I literally started learning this a week ago! Does anybody fancy explaining to me how you use the chi-squared to test the goodness of fit and how you work out the expected frequencies and that? That's the only part I don't understand!

Haven't you done the Chi-square test on S2?

Well, here's how I set out the question.

H0: The model fits the data.
H0: The model does not fit the data.

Work out expected, sometimes they give you these, and sometimes they tell you to work them out, like probabilities for X~Poisson(4.2), X~N(54,6^2), X~B(5,0.1)
Then work out contributions using [(O-E)^2]/E
Remember, if any expected frequencies are less than 5, group the data.

X% significance level
Test statistics X^2= sum of contributions
There n classes, m parameters were estimated (for Poisson, Binomial you estimate the mean so m=1, for Normal you estimate the mean and variance so m=2, if you didn't estimate any parameters, then m=0).
So degrees of freedom v = n-m-1
At the X% significance level with v=n-m-1, criitical value=...

Since X^2<critical value, result is significant.
There is insufficient evidence to reject H0.
The evidence suggests that the model fits the data.
OR
Since X^2>critical value, result is not significant.
There is sufficient evidence to reject H0.
The evidence suggests that the model does not fit the data.

[Sharma93]

(Original post by lubamira)
I've left doing work for this exam way too late! (I had C3 and FP2 today so I needed to concentrate on those) However, I was just wondering if anyone knew anything about interpolation? I've only just heard about it today!

Interpolation? Wasn't that S2?

[james20207]

(Original post by albus)
Haven't you done the Chi-square test on S2?

Well, here's how I set out the question.

H0: The model fits the data.
H0: The model does not fit the data.

Work out expected, sometimes they give you these, and sometimes they tell you to work them out, like probabilities for X~Poisson(4.2), X~N(54,6^2), X~B(5,0.1)
Then work out contributions using [(O-E)^2]/E
Remember, if any expected frequencies are less than 5, group the data.

X% significance level
Test statistics X^2= sum of contributions
There n classes, m parameters were estimated (for Poisson, Binomial you estimate the mean so m=1, for Normal you estimate the mean and variance so m=2, if you didn't estimate any parameters, then m=0).
So degrees of freedom v = n-m-1
At the X% significance level with v=n-m-1, criitical value=...

Since X^2<critical value, result is significant.
There is insufficient evidence to reject H0.
The evidence suggests that the model fits the data.
OR
Since X^2>critical value, result is not significant.
There is sufficient evidence to reject H0.
The evidence suggests that the model does not fit the data.

Thanks so much for taking the time to explain it thoroughly. You have saved me about 10 marks if a big question on this comes up. How are you feeling for this then?

[lubamira]

(Original post by Sharma93)
Interpolation? Wasn't that S2?

I don't know, I've not been taught it at all! Someone mentioned it today and it freaked me out!

[albus]

(Original post by lubamira)
I don't know, I've not been taught it at all! Someone mentioned it today and it freaked me out!

All the stuff on S3 is
Continuous Random Variables
Expectation Algebra
Sampling
Confidence intervals
t test
Wilcoxon Test
Chi-square Goodness of Fit Test.

[Benniboi1]

(Original post by albus)
All the stuff on S3 is
Continuous Random Variables
Expectation Algebra
Sampling
Confidence intervals
t test
Wilcoxon Test
Chi-square Goodness of Fit Test.

For the Chi-squared tests, how do you know what the value of v is for the different tests? It's really confused me! thanks

[Maths-geek]

I thought that was such a nice paper, nothing too tricky really. What did you guys get for v when finding the critical value for the chi-squared, i put v=2 but a couple of my friends put v=3?

[alexbumpoo]

(Original post by Maths-geek)
I thought that was such a nice paper, nothing too tricky really. What did you guys get for v when finding the critical value for the chi-squared, i put v=2 but a couple of my friends put v=3?

how many classes did you group together

[justanotherposter]

What a great paper! Only problems I had was with finding the CDF but that's only a couple of marks.
V was 3 by the way, you only subtract one from the degrees of freedom because you aren't estimating the mean, it was found from the data, you only subtract an extra 1 if you estimate it from nowhere.

[Maths-geek]

(Original post by alexbumpoo)
how many classes did you group together

The classes I ended up with were 0, 1, 2 and 3 or more, so 4 classes

[Wolvesfc94]

wasn't the mean estimated from a sample though? sorry if i'm being stupid....

what did you get for F(t)?

[Maths-geek]

(Original post by Wolvesfc94)
wasn't the mean estimated from a sample though? sorry if i'm being stupid....

what did you get for F(t)?

I thought the mean was estimated as well

For F(t) I got something like 1-e^-(lambda(t))

[albus]

(Original post by Maths-geek)
I thought that was such a nice paper, nothing too tricky really. What did you guys get for v when finding the critical value for the chi-squared, i put v=2 but a couple of my friends put v=3?

I got v=2 aswell, you group the last 3 groups together.
So there were 4 classes, and the Poisson mean was estimated, so that was 1 parameter. so v=4-1-1=2