+ All Categories
Home > Documents > LECTURE_02.pptx

LECTURE_02.pptx

Date post: 07-Aug-2018
Category:
Upload: cengiz-kaya
View: 212 times
Download: 0 times
Share this document with a friend

of 17

Transcript
  • 8/21/2019 LECTURE_02.pptx

    1/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis TestingIn deriving Bayes decision rule, we assumed that we know both the a priori

    probabilities 0 and 1 as well as the likelihoods )|( 0Hyp and )|( 1Hyp . Thismeans that we have both the knowledge of the mechanism generating the state of the

    nature and the mechanism affecting our observations (our measurements about the

    state of nature). It is, however, possible that we may not have access to all the

    information. For example, we may not know the a prioriprobabilities. In such a case,

    the Bayes decision rule is not a good rule since it can only be derived for a given a

    prioriprobability.An alternative to the Byes hypothesis testing, in this case, is the Minimax

    Hypothesis testing. The minimax decision rule minimizes the maximum possible risk,

    i.e., it minimizes,

    )(),(max 10 RR over all .

    Lets look at),( 0 r , i.e., the overall risk for a decision rule when the a priori

    probability is]1,0[0 . It can be written as:

    )()1()(),( 10000 RRr

  • 8/21/2019 LECTURE_02.pptx

    2/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis Testing

    Note that, for a given decision rule , as 0 varies from 0 to 1, ),( 0 r goes

    linearly from ),0()(1 rR to ),1()(0 rR . Therefore, for a given decision rule

    the maximum value of ),( 0 r as 0 varies over the interval ]1,0[ occurs either at

    00 or 10 and it is )}(),(max{ 10 RR .

    So minimizing )}(),(max{ 10 RR is equivalent to minimizing ),(max 010 0

    r

    .

    Thus, the minimax decision rule is:),(maxmin 0

    10 0

    r

    Let0

    denote the optimum (Bayes) decision rule for the a priori probability 0 .

    Denote the corresponding minimum Bayes risk by )( 0V , i.e., ),()( 00 rV .

    It is easy to show that )( 0V is a continuous concave function of 0 for

    ]1,[0 o and has the end points 11)0( CV and 00)1( CV .

  • 8/21/2019 LECTURE_02.pptx

    3/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis Testing

    The following figure shows a typical graph of )( 0V and),( 0 r . Lets draw a tangent to )( 0V parallel to ),( 0 r .

    Denote this line by ),(00

    r .

  • 8/21/2019 LECTURE_02.pptx

    4/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis TestingSince ),(

    00 r lies entirely below the line ),( 0 r , it has a lower maximum

    compared to ),( 0 r . Also note that since it touches )( 0V at 00 then 0 is the

    minimum risk (Bayes) rule for a prioriprobability 0 . Since for any ]1,0[0 we can

    draw a tangent to )( 0V and find the minimum risk rule as a Bayes rule, it is clear that

    the minimax decision rule is the Bayes rule for the value of 0 that maximizes )( 0V .

    Denoting this point by L , we note that point,

    )()()}(),(max{ 1010 LLLL RRRR

  • 8/21/2019 LECTURE_02.pptx

    5/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis Testing

    Proposition: The Minimax Test

    Let L be the a prioriprobability that maximizes )( 0V and such that either 0L , or 1L , or

    ),()( 10 LL RR

    thenL

    is a minimax rule.

  • 8/21/2019 LECTURE_02.pptx

    6/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis Testing

    Proof

    Let ),()( 10 LL RR then for any 0 we have,

    ),,(),(),(minmax 0010 0

    LL rrr L

    So, we have

    ),,(maxmin),(max),(minmax 010

    010

    010 000

    rrrL

    Also, for each we have).,(minmax),(max 0

    100

    10 00

    rr

    This implies that,

    ).,(minmax),(maxmin 010

    010 00

    rr

    Combining the two inequalities, we get,

    .),(minmax),(maxmin 010

    010 00

    rr

    Therefore,

    .),(maxmin),( 010 0

    rr LL

    That is,L

    is the minimax rule.

  • 8/21/2019 LECTURE_02.pptx

    7/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis TestingDiscussion

    By definition: ),()(0

    00 rV . So, for every ]1,0[0 , we have )(),( 00

    0 Vr and

    ).(),( 00

    0 Vr Since ),(

    00 r , as a function of 0 is a straight line, it has to be tangent to

    )( 0V at .00 If )( 0V is differentiable at 0 , we have,

    ).()(/),()(000 10000

    RRddrV

    Now consider the case that )( 0V has an interior maximum but is not differentiable at that point. In

    this case we define two decision rules 00lim LL

    and .lim00 LL

    The critical regions for these two decision rules are,

    ),|()()|())(1(|{01000101111

    HypCCHypCCyLL

    and

    ),|()()|())(1(|{01000101111

    HypCCHypCCyLL

    Take a number ]1,0[q and devise a decision rule~

    L that uses the decision rule

    L with probability

    q and uses L

    with probability q1 . It means that it decides1

    H if 1

    y , decides0

    H if cy )(1

    and

    decides 1H with probability q ify is on the boundary of 1 .

  • 8/21/2019 LECTURE_02.pptx

    8/17

    ELEC6111: Detection and Estimation Theory

    Minimax Hypothesis TestingDiscussion

    Note that the Bayes risk is not a function ofq , so )(),(

    ~

    LL Vr L but the conditional risksdepend on q ,

    ).()1()()(~

    LLL jjj

    RqqRR

    To achieve )()(~

    1

    ~

    0 LLRR , we need to choose,

    .

    )()()()(

    )()(

    1010

    10

    LLLL

    LL

    RRRR

    RRq

    Note that )()()( 00 LL

    RRV L , so we have:

    .)()(

    )(

    LL

    L

    VV

    Vq

    This is called a randomized decision rule.

  • 8/21/2019 LECTURE_02.pptx

    9/17

    ELEC6111: Detection and Estimation Theory

    Example: Measurement with Gaussian Error

    Consider the measurement with Gaussian error with unifom costs

    The function )( 0V can be written as,

    )),(1)(1()()( 100

    00

    QQV

    With

    .2

    )1

    log( 10

    0

    0

    01

    2

    We can find the rule making conditional risks )(0 R and )(1 R equal by letting,

    ))(1()( 10

    QQ

    and solving for .

  • 8/21/2019 LECTURE_02.pptx

    10/17

    ELEC6111: Detection and Estimation Theory

    Example: Measurement with Gaussian Error

    We can solve this by inspection and get:

    .2

    10

    L

    So, the minimax decision rule is:

    .2/)(0

    2/)(1)(

    10

    10

    yif

    yify

    L

    Conditional risks for measurement with Gaussian error

  • 8/21/2019 LECTURE_02.pptx

    11/17

    ELEC6111: Detection and Estimation TheoryNeyman-Pearson Hypothesis Testing

    In Bayes hypothesis testing as well as minimax, we are concerned with the average risk, i.e., the

    conditional risk averaged over the two hypotheses. Neyman-Pearson test, on the other hand, recognizes theasymmetry between the two hypotheses. It tries to minimize one of the two conditional risks with the other

    conditional risk fixed (or bounded).

    In testing the two hypotheses 0H and 1H , the following situations may arise:

    0H is true but 1H is decided. This is called a type I error or afalse alarm. This comes from radarapplication where 0H represents no target and 1H is the case of target present. The probability

    of this event is calledfalse alarm probability orfalse alarm rate and is denoted as )(F

    P

    1H is true but 0H is decided. This is called a type II error or a miss. The probability of this event iscalled miss probability and is denoted as )(MP

    0H is true and 0H is decided. Probability of this event is )(1 FP . 1H is true and 1H is decided. This case represents a detection. The detection probabilityis

    )(1)( MD PP .

    In testing 0H versus 1H , one has to tradeoff between the probabilities of two types of errors. Neyman-

    Pearson criterion makes this tradeoff by bounding the probability of false alarm and minimizing miss

    probability subject to this constraint, i.e., the Neyman-Pearson test is,

    )(max

    DP subject to ,)( FP

    where is the bound on false alarm rate. It is called the levelof the test.

  • 8/21/2019 LECTURE_02.pptx

    12/17

    ELEC6111: Detection and Estimation TheoryNeyman-Pearson Hypothesis Testing

    For obtaining a general solution to the Neyman-Pearson test, we need to define a randomized decisionrule. We define the randomized test,

    L

    L

    L

    yLif

    yLifq

    yLif

    yL

    )(0

    )(

    )(1

    )(~

    where L is the threshold corresponding to L .

    While in a non-randomized rule, )(y gives the decision, in a randomized rule, )(

    ~

    yL gives the probability ofdecision.

    Then we have,

    ,)|()()}({)( 0~~

    0

    ~

    dyHypyYEPF

    where {.}0E is expectation under hypothesis 0H . Also,

    .)|()()}({)(1

    ~~

    1

    ~

    dyHypyYEPD

  • 8/21/2019 LECTURE_02.pptx

    13/17

    ELEC6111: Detection and Estimation TheoryNeyman-Pearson Lemma

    Consider a hypothesis pair 0H and 1H :00 ~: PYH

    Versus

    11 ~: PYH

    where jP has density )|()( jj Hypyp for 1,0j . For 0 , the following statements are true:

    1. Optimality: Let ~be any decision rule satisfying .)( ~ FP Let ~be any decision rule of the form)(

    ),|()|(0

    )|()|()()|()|(1

    01

    01

    01~

    A

    HypHypif

    HypHypifyHypHypif

    where 0 and 1)(0 y are such that .)(~

    FP Then ).()(~~

    DD PP

    This means that any size- decision rule of form (A) is Neyman-Pearson rule.

    2. Existence: For any )1,0( there is a decision rule, NP~ , of form (A) with 0)( y for which.)(

    ~

    NPFP

    3. Uniqueness: Suppose that is any Neyman-Pearson rule of size- for 0H versus 1H . Then must be ofthe form (A).

  • 8/21/2019 LECTURE_02.pptx

    14/17

    ELEC6111: Detection and Estimation TheoryNeyman-Pearson Lemma (Proof)

    1.Not that, by definition, we always have 0)]|()|()][()([ 01

    ~~

    HypHypyy (why?)So, we have,

    .0)]|()|()][()([ 01

    ~~

    dyHypHypyy Expanding the above expression, we get,

    .)|()()|()()|()()|()( 0

    ~

    0

    ~

    1

    ~

    1

    ~

    dyHypydyHypydyHypydyHypy

    Applying the expressions for the detection probabilityandfalse alarm rate, we have:

    .0)]([)]()([)()(~~~~~ FFFDD PPPPP

    2. Let 0 be the smallest number such that (look at the Figure in next slide): )]|()|([ 0010 HYpHYpP .

    Then if )]|()|([ 0010 HYpHYpP , choose,

    )]|()|([

    )]|()|([

    0010

    00100

    HYpHYpP

    HYpHYpP

    Otherwise, choose 0 arbitrarily. Consider a Neyman-Pearson decision rule, NP~

    , with 0 and 0)( y . For this

    decision rule, the false alarm arte is,

    )]|()|([)]|()|([}{)( 001000010~

    0

    ~

    HYpHYpPHYpHYpPEP NPNPF .

  • 8/21/2019 LECTURE_02.pptx

    15/17

    ELEC6111: Detection and Estimation TheoryNeyman-Pearson Lemma (Proof)

    .

    2. See the text.

  • 8/21/2019 LECTURE_02.pptx

    16/17

    ELEC6111: Detection and Estimation TheoryNeyman-Pearson Lemma (Example): Measurement with Gaussian Error

    For this problem, we have,

    ),(1)()())([)]|()|([ 00010

    QYPyLPHYpHYpP

    where .2

    )log( 10

    01

    2

    Any value of can be achieved by choosing,

    .)1()( 01

    0

    1

    0

    Q

    Since 0)( 0 YP , the choice of 0 is arbitrary and we can choose 10 . So, we have

    .0

    1)(

    0

    0~

    yif

    yifyNP

  • 8/21/2019 LECTURE_02.pptx

    17/17

    ELEC6111: Detection and Estimation TheoryNeyman-Pearson Lemma (Example): Measurement with Gaussian Error

    The detection probability for NP

    ~

    is dQQQQQYPYEP NPNPD

    )()()()()}({)( 10111001

    ~

    1

    ~


Recommended