+ Reply to Thread
Results 1 to 5 of 5

Thread: the MLE question. what is the MLE if the variance is heteroskedasticity

  1. #1
    Points: 1,718, Level: 24
    Level completed: 18%, Points required for next Level: 82

    Location
    korea
    Posts
    14
    Thanks
    0
    Thanked 0 Times in 0 Posts

    the MLE question. what is the MLE if the variance is heteroskedasticity




    The normal distribution of MLE is start {x1,,,,,xn} random sample from N(u,σ^2)

    This family of distributions has two parameters: θ = (μ, σ), so we maximize the likelihood, , over both parameters simultaneously using logarithm...

    I know this. but what if the normal distribution's {x1,,,,,xn} random sample from N(u,σi^2)

    I mean the variance is not homoscedasticity. now the variance is heteroskedasticity...

    I atteched the image. My questions is that What is the MLE if variance change σ^2 to σi^2.
    σ^2 - homoscedasticity,
    σi^2- variance is heteroskedasticity.
    So

    f(x/u, σi^2) = 1/sqrt(2pi)σi exp((x-u)^2/2σi^2
    I don't know.... How can I solved...
    Attached Images  
    Last edited by wawataiji; 03-22-2013 at 03:20 AM.

  2. #2
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: the MLE question. what is the MLE if the variance is heteroskedasticity

    Your notation and phrasing make it difficult to understand what you're asking. Are you saying each observation has a different unknown variance?
    I don't have emotions and sometimes that makes me very sad.

  3. #3
    Points: 1,718, Level: 24
    Level completed: 18%, Points required for next Level: 82

    Location
    korea
    Posts
    14
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: the MLE question. what is the MLE if the variance is heteroskedasticity

    I atteched the image. My questions that What is the MLE if the variance is change σ^2 to σi^2.
    σ^2 - homoscedasticity,
    σi^2- variance is heteroskedasticity.
    So f(x/u, σi^2) = 1/sqrt(2pi)σi exp((x-u)^2/2σi^2

  4. #4
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: the MLE question. what is the MLE if the variance is heteroskedasticity

    Your terminology doesn't make sense. Once again I am asking are you saying that every observation [/math]x_i[/math] has their own variance term \sigma_i^2? Heteroskedasticity means that the variance isn't constant but typically if we model that we either have repeated observations (multiple x's take on the same value) or we model the variance as a function of the mean or something...

    If you want to estimate the overall mean and estimate a completely different variance for each observation... then you're out of luck because you have one more parameter than you do observations and there isn't a unique solution.
    I don't have emotions and sometimes that makes me very sad.

  5. #5
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: the MLE question. what is the MLE if the variance is heteroskedasticity


    Your terminology doesn't make sense. Once again I am asking are you saying that every observation [/math]x_i[/math] has their own variance term \sigma_i^2? Heteroskedasticity means that the variance isn't constant but typically if we model that we either have repeated observations (multiple x's take on the same value) or we model the variance as a function of the mean or something...

    If you want to estimate the overall mean and estimate a completely different variance for each observation... then you're out of luck because you have one more parameter than you do observations and there isn't a unique solution.
    I don't have emotions and sometimes that makes me very sad.

+ Reply to Thread

           




Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats