Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie
Hi there,
There is an issue with role permissions that is being worked on at the moment.
If you are having trouble with access or permissions on regional forums please post here to get access: https://www.boards.ie/discussion/2058365403/you-do-not-have-permission-for-that#latest

Mathematical Statistics question

  • 05-05-2009 9:00am
    #1
    Closed Accounts Posts: 48


    Here it is,

    Let the General Linear Model be represented by

    y= Xß + u, u ~ N(0, (σ^2)I)

    where X is n x p non-stochastic and of rank p and let ß = (X^t)y / [(X^t)X] be the OLS estimate of ß. Let c be a p x 1 vector, determine the distribution of (c^t)(ß), and hence devise a test statistic to test the hypothesis that (c^t)ß = d, for any number d.

    If anyone could help me out with this it would be unreal. Totally lost.


Comments

  • Closed Accounts Posts: 6,151 ✭✭✭Thomas_S_Hunterson


    What exactly is it you don't understand? Is it the linear algebra? devising the test statistic? or just the whole question?

    It's just a hypothesis test for a point estimate of y.

    c is just the vector for an observation.
    ß is the vector of coeficients

    Multiplying them gives the estimate for y.

    Have you access to a library? If so I'd recommend Applied Linear Statistical Models 5th ed, by Kutner, Nachtsheim, Neter & Li. IIRC chapter 5 covers the linear algebra of the models.


  • Closed Accounts Posts: 48 qwertykeyboards


    thanks for the help. pretty much lost with all of it. im lookin through my lecturers notes but they're too long and cant seem to start the question. in library now just grabbed the 3rd edition of that book you said, its the latest copy here. gonna try read some of it and see if i can make sense of it. thanks


  • Closed Accounts Posts: 48 qwertykeyboards


    read through all these notes now, is it common paractice to begin with 3 assumptions: 1. E[ui] = 0, 2. V[ui] = σ^2, 3. Cov[ui, uj] = 0

    is that where i should start to do my test statistic?


  • Closed Accounts Posts: 6,151 ✭✭✭Thomas_S_Hunterson


    read through all these notes now, is it common paractice to begin with 3 assumptions: 1. E[ui] = 0, 2. V[ui] = σ^2, 3. Cov[ui, uj] = 0

    Yea, all these assumptions are given in the question, albeit phrased differently:
    u ~ N(0, (σ^2)I)

    What it's saying is that observations will be normally distributed around the predicted value (for given X) with variance σ^2.


  • Closed Accounts Posts: 2 cbhoy


    Just you wait til Gary Keogh sees this!:mad:


  • Advertisement
  • Registered Users, Registered Users 2 Posts: 1,152 ✭✭✭carlowboy


    Here it is,

    Let the General Linear Model be represented by

    y= Xß + u, u ~ N(0, (σ^2)I)

    where X is n x p non-stochastic and of rank p and let ß = (X^t)y / [(X^t)X] be the OLS estimate of ß. Let c be a p x 1 vector, determine the distribution of (c^t)(ß), and hence devise a test statistic to test the hypothesis that (c^t)ß = d, for any number d.

    If anyone could help me out with this it would be unreal. Totally lost.

    Is that for the assignment?


Advertisement