Scribd Logo
Upload

Welcome to Scribd!

  • 61 views

    Properties of Distribution Function of A Random Vector

    Uploaded by

    pankaj kumar

    Copyright:

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

    Properties of Distribution Function of A Random Vector

    Uploaded by

    pankaj kumar
    61 views12 pages

    Original Title

    Module_20_0.pdf

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    61 views12 pages

    Properties of Distribution Function of A Random Vector

    Uploaded by

    pankaj kumar

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 12

    Module 20

    Properties of Distribution Function of a Random Vector

    () Module 20 Properties of Distribution Function of a Random Vector 1 / 11


    X: a given p-dimensional r.v. defined on a probability space
    (Ω, P(Ω), P) with d.f. FX (·);

    FX (x) = P({X ≤ x})


    = PX ((−∞, x]), x ∈ Rp ,

    where PX (·) is the probability function induced by X ;

    For any p-dimensional rectangle (a, b] (a < b), we know that


    p
    X X
    P({X ∈ (a, b]}) = (−1)k FX (z).
    k=0 z∈∆k,p ((a,b])

    () Module 20 Properties of Distribution Function of a Random Vector 1 / 11


    Result 1:

    Let X = (X1 , . . . , Xp ) be a p-dimensional r.v. with d.f. FX (·). Then


    (a) lim F (x , . . . , xp )
    xi →∞ X 1
    = 1; (iterated limit)
    i=1,...,p
    (b) For each fixed i ∈ {1, . . . , p} , lim FX (x1 , . . . , xp ) = 0;
    xi →−∞
    (c) FX (x1 , . . . , xp ) is right continuous in each argument, keeping other
    arguments fixed;
    (d) For each rectangle (a, b] ⊆ Rp (a, b ∈ Rp , a < b)
    p
    X X
    (−1)k FX (z) ≥ 0.
    k=0 z∈∆k,p ((a,b])

    () Module 20 Properties of Distribution Function of a Random Vector 2 / 11


    Proof:

    For p = 2
    (a)
    h i
    lim lim FX1 ,X2 (m, n) = lim lim FX1 ,X2 (m, n)
    m→∞ n→∞ m→∞ n→∞

    = lim FX1 (m)


    m→∞

    = 1.

    () Module 20 Properties of Distribution Function of a Random Vector 3 / 11


    (b) For x2 ∈ R,

    lim FX1 ,X2 (x1 , x2 ) = lim FX1 ,X2 (−m, x2 )


    x1 →−∞ m→∞
    = lim P ({X1 ≤ −m, X2 ≤ x2 })
    m→∞
    ∞
    !
    \
    = P {X1 ≤ −m, X2 ≤ x2 }
    m=1
    = P(φ)
    = 0.

    Similarly, for x1 ∈ R,

    lim FX1 ,X2 (x1 , x2 ) = 0.


    x2 →−∞

    () Module 20 Properties of Distribution Function of a Random Vector 4 / 11


    (c) For (x1 , x2 ) ∈ R2
     
    1 1
    lim FX1 ,X2 (x1 + , x2 ) = lim P X1 ≤ x1 + , X2 ≤ x2
    n→∞ n n→∞ n
    ∞  !
    \ 1
    = P X1 ≤ x1 + , X2 ≤ x2
    n
    n=1
    = P ({X1 ≤ x1 , X2 ≤ x2 })
    = FX1 ,X2 (x1 , x2 ),

    implying that FX1 ,X2 (·) is right continuous in first argument. Similarly
    FX1 ,X2 (·) is right continuous in second argument.

    () Module 20 Properties of Distribution Function of a Random Vector 5 / 11


    (d) Let −∞ < a1 < b1 < ∞ and −∞ < a2 < b2 < ∞. Then
    2
    X X
    (−1)k FX (z) = FX (b1 , b2 ) − FX (a1 , b2 )
    k=0 z∈∆k,2 ((a,b])
    −FX (b1 , a2 ) + FX (a1 , a2 )
    = P ({a1 < X1 ≤ b1 , a2 < X2 ≤ b2 })
    ≥ 0.

    () Module 20 Properties of Distribution Function of a Random Vector 6 / 11


    Remark 1:
    (a) Let FX (·) be the d.f. of a p-dimensional r.v. X . For h > 0 and
    (a1 , a2 , . . . , ap ) ∈ Rp
    FX (a1 + h, a2 , . . . , ap ) − FX (a1 , a2 , . . . , ap )
    = P({X1 ≤ a1 + h, Xi ≤ ai , i = 2, . . . , p})
    −P({X1 ≤ a1 , Xi ≤ ai , i = 2, . . . , p})
    = P({a1 < X1 ≤ a1 + h, Xi ≤ ai , i = 2, . . . , p})
    ≥ 0.
    It follows that the d.f. of a random vector is non-decreasing in each
    argument when other arguments are kept fixed.
    (b) For p = 1, condition (d) of above theorem is equivalent to
    P ({a < X ≤ b}) ≥ 0, ∀ − ∞ < a < b < ∞
    ⇔ FX (b) ≥ FX (a) , ∀ − ∞ < a < b < ∞,
    i.e., FX (·) is non-decreasing.
    () Module 20 Properties of Distribution Function of a Random Vector 7 / 11
    Result 2:

    If a function G : R → R satisfies properties mentioned in (a)-(d) of above


    result then there exists a probability space (Ω, P(Ω), P) and a r.v.
    X = (X1 , . . . , Xp ) on Ω such that G (·) is the d.f. of X .

    Remark 3: Clearly the d.f.

    FX (x) = P ({X ≤ x})


    = PX ((−∞, x]) , x ∈ Rp

    is determined by the induced probability function PX (·). Conversely, it can


    be shown that, d.f. determines the induced probability function PX (·)
    uniquely. This suggests that to study the induced probability function
    PX (·) it suffices to study the d.f. FX (·).

    () Module 20 Properties of Distribution Function of a Random Vector 8 / 11


    Take Home Problem

    Let F : R → R be given by
    
    1, if x + 2y ≥ 1
    F (x, y ) = .
    0, if x + 2y < 1

    Does F (·) defines a d.f. of some random vector?

    () Module 20 Properties of Distribution Function of a Random Vector 9 / 11


    Abstract of Next Module

    We will introduce the notion of independence of random variables.

    () Module 20 Properties of Distribution Function of a Random Vector 10 / 11


    Thank you for your patience

    () Module 20 Properties of Distribution Function of a Random Vector 11 / 11

    You might also like