Board index › Topical Discussion › Computers, Math, Science, and Technology

Statistics problem

Page 1 of 1 [ 3 posts ] 

Previous topic | Next topic

Kurgan
Veteran
Veteran

User avatar

Joined: 6 Apr 2012
Age: 35
Gender: Male
Posts: 4,132
Location: Scandinavia

18 Sep 2014, 6:49 pm

I haven't had statistics in almost 3 years, and I've encountered a problem that goes as follows:

Image

The book used in this course is completely rubbish. So how do I solve this?


_________________
“He who controls the spice controls the universe.”


NGC6205
Hummingbird
Hummingbird

User avatar

Joined: 7 Jun 2014
Age: 28
Gender: Male
Posts: 20
Location: 16h41m41.24s,+36°27′35.5″

18 Sep 2014, 8:49 pm

I am presuming this N(1,2) notation refers to mu=1 and sigma^2=2, which appears consistent based on the fact the problem works.

First for the mean:
You will have t terms of 1 divided by t, so therefore the mean of the running mean will be 1.

Next for the variance:
Since variance of a sum of random variables in the sum of the individual variances, the summation alone will have a variance of 2*t. However, multiplying a random variable by a constant linearly changes the standard deviation, dividing the distribution by t will actually divide the variance by t^2. Therefore the net result is 2/t.

Thus the distribution is equal to N(1,2/t).



Kurgan
Veteran
Veteran

User avatar

Joined: 6 Apr 2012
Age: 35
Gender: Male
Posts: 4,132
Location: Scandinavia

19 Sep 2014, 7:39 am

NGC6205 wrote:
I am presuming this N(1,2) notation refers to mu=1 and sigma^2=2, which appears consistent based on the fact the problem works.

First for the mean:
You will have t terms of 1 divided by t, so therefore the mean of the running mean will be 1.


Yes, but there are t terms of X_i, which is where the problems start showing up. :) Edit: I think I know what you mean. Thanks.


_________________
“He who controls the spice controls the universe.”