Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 15 views

    Lecture13 14

    Uploaded by

    Anonymous T02GVGzB
    This document summarizes key points from a statistics lecture, including: 1) The sample variance S^2 of a normal distribution with parameters m and σ^2 follows a chi-squared distribution. This is used to obtain an unbiased estimator of σ^2 and compute its variance and Cramer-Rao lower bound. 2) The Cramer-Rao lower bound gives the minimum possible variance for any unbiased estimator based on n observations. Estimators that achieve this bound are efficient and have variance that decreases to zero as n approaches infinity. 3) Properties of the maximum likelihood estimator are analyzed for distributions with specific parameters, including whether it is unbiased and efficient. 4) A

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Lecture13 14

    Uploaded by

    Anonymous T02GVGzB
    15 views2 pages
    This document summarizes key points from a statistics lecture, including: 1) The sample variance S^2 of a normal distribution with parameters m and σ^2 follows a chi-squared distribution. This is used to obtain an unbiased estimator of σ^2 and compute its variance and Cramer-Rao lower bound. 2) The Cramer-Rao lower bound gives the minimum possible variance for any unbiased estimator based on n observations. Estimators that achieve this bound are efficient and have variance that decreases to zero as n approaches infinity. 3) Properties of the maximum likelihood estimator are analyzed for distributions with specific parameters, including whether it is unbiased and efficient. 4) A

    Original Description:

    estadistica fisica

    Original Title

    lecture13-14

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document summarizes key points from a statistics lecture, including: 1) The sample variance S^2 of a normal distribution with parameters m and σ^2 follows a chi-squared distribution. This is used to obtain an unbiased estimator of σ^2 and compute its variance and Cramer-Rao lower bound. 2) The Cramer-Rao lower bound gives the minimum possible variance for any unbiased estimator based on n observations. Estimators that achieve this bound are efficient and have variance that decreases to zero as n approaches infinity. 3) Properties of the maximum likelihood estimator are analyzed for distributions with specific parameters, including whether it is unbiased and efficient. 4) A

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    15 views2 pages

    Lecture13 14

    Uploaded by

    Anonymous T02GVGzB
    This document summarizes key points from a statistics lecture, including: 1) The sample variance S^2 of a normal distribution with parameters m and σ^2 follows a chi-squared distribution. This is used to obtain an unbiased estimator of σ^2 and compute its variance and Cramer-Rao lower bound. 2) The Cramer-Rao lower bound gives the minimum possible variance for any unbiased estimator based on n observations. Estimators that achieve this bound are efficient and have variance that decreases to zero as n approaches infinity. 3) Properties of the maximum likelihood estimator are analyzed for distributions with specific parameters, including whether it is unbiased and efficient. 4) A

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 2

    Lecture March, 27, 2012.

    1. When we take a sample of n independent observations X1 , . . . , Xn from


    P
    a N (m, 2 ) population, the sample variance S 2 = n1 ni=1 (Xi X)2 can
    be shown to verify:
    nS 2
    2n1 .
    2

    Using this fact (and remembering that a 2n distribution has mean n


    and variance 2n),
    (a) Obtain an unbiased estimator
    2 of 2 .
    (b) Compute its variance.
    (c) Compute the Cramer-Rao lower bound for the variance of unbiased
    estimators of 2 .
    (d) Is
    2 efficient?
    2. Write the expression of the Cramer-Rao lower bound for the variance
    of unbiased estimators of a parameter based on n observations, under
    regularity conditions. Then answer the following questions:
    (a) If an unbiased estimator attains the Cramer-Rao lower bound, can
    we assert that its variance decreases to zero as n ? At which
    rate?
    (b) Is such an estimator consistent?
    3. Let X be distributed in (0, ) with density fX (x) = 3x2 /3 . Consider
    the estimator = 43 X. Is it unbiased? What is its variance? Can we
    use the Cramer-Rao lower bound to check its efficiency?
    4. According to a well known model (the Hardy-Weinberg equilibrium),
    the offspring of a generation in which two characters have frequencies p
    and q = 1 p are distributed in three classes, with probabilities p2 , 2pq
    and q 2 . If we have a sample in which those three classes have absolute
    frequencies 158, 475 and 362, write the equation giving the MLE of p.
    5. X is distributed as binary, with parameter p. We know that

    1
    3

    (a) What is the MLE of p based on one single observation?


    (b) Is it unbiased?
    (c) What is its mean squared error if the true p is 0.5?
    1

    p 23 .

    (d) Is the MLE optimal? (Hint: Consider the rather silly estimator p = 0.5, irrespective of what is the value of the single
    observation. This simple example is meant to show that the MLE
    need not be very good if the sample is small.)
    6. A plausible estimator of m in a N (m, 2 ) distribution would be the
    sample median, since we know in that distribution the mean m and the
    median take the same value. It can be shown that, asymptotically, the
    sample median M e verifies:

    n(M e m) N (0, 2 = (0.5/fX (m))2


    What is the efficiency of M e?
    You may find similar and additional problems in many books, such as [?],
    [?], [?], [?] and [?] (the last, at a higher level than this course).

    You might also like