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

Welcome to Scribd!

  • 42 views

    EXERCISE 6: Discriminant Analysis

    Uploaded by

    Lian Garcia
    1. A chi-square test was conducted to test if the variance-covariance matrices of two groups (college and high school football players) were significantly different. The p-value was greater than 0.05, so the null hypothesis that the matrices were not different was not rejected. 2. Linear discriminant functions were computed to classify players into groups based on head measurements. Cross-validation estimated a misclassification rate of 30%. 3. An F-test showed the linear discriminant function had discriminating ability, as the calculated F value exceeded the critical value.

    Copyright:

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

    EXERCISE 6: Discriminant Analysis

    Uploaded by

    Lian Garcia
    42 views4 pages
    1. A chi-square test was conducted to test if the variance-covariance matrices of two groups (college and high school football players) were significantly different. The p-value was greater than 0.05, so the null hypothesis that the matrices were not different was not rejected. 2. Linear discriminant functions were computed to classify players into groups based on head measurements. Cross-validation estimated a misclassification rate of 30%. 3. An F-test showed the linear discriminant function had discriminating ability, as the calculated F value exceeded the critical value.

    Original Title

    EXER6_Calubayan_Garcia

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    1. A chi-square test was conducted to test if the variance-covariance matrices of two groups (college and high school football players) were significantly different. The p-value was greater than 0.05, so the null hypothesis that the matrices were not different was not rejected. 2. Linear discriminant functions were computed to classify players into groups based on head measurements. Cross-validation estimated a misclassification rate of 30%. 3. An F-test showed the linear discriminant function had discriminating ability, as the calculated F value exceeded the critical value.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    42 views4 pages

    EXERCISE 6: Discriminant Analysis

    Uploaded by

    Lian Garcia
    1. A chi-square test was conducted to test if the variance-covariance matrices of two groups (college and high school football players) were significantly different. The p-value was greater than 0.05, so the null hypothesis that the matrices were not different was not rejected. 2. Linear discriminant functions were computed to classify players into groups based on head measurements. Cross-validation estimated a misclassification rate of 30%. 3. An F-test showed the linear discriminant function had discriminating ability, as the calculated F value exceeded the critical value.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 4

    Calubayan, George Emmanuel STAT 175

    V-3L
    Garcia, Lian Louis

    EXERCISE 6: Discriminant Analysis

    1. H​o​: The variance-covariance matrices of the two groups are not significantly different.
    H​a​: The variance-covariance matrices of the two groups are significantly different.

    Test procedure: Chi-Square Test


    ​ ​p​-value​ < α = 0.05​; Otherwise, fail to reject H​o.​
    Decision Rule : Reject H​o if

    SAS Output:

    p​-value: 0.5684
    Decision: Since ​p​-value = 0.5684 ​> α = 0.05​, we ​fail to reject H​o.​
    Conclusion: At 5% level of significance, there is sufficient evidence to say that the
    variance-covariance matrices of the two groups are ​not significantly different​.

    Assuming equal prior probabilities, the linear discriminant function will be used.
    2. The computed coefficients are

    Variable College HS

    Constant -1122 -1098

    WDIM 21.15704 20.7271

    CIRCUM 37.85851 37.87838

    FBEYE -24.60763 -24.06528

    JAW 15.75632 13.36753

    The estimated LDFs are:


    z1​ =​ -​ 1122 + 21.15704​WDIM​ + 37.85851​CIRCUM​ - 24.60763​FBEYE​ +
    15.75632​JAW
    z​2 =
    ​ - ​ 1098 + 20.72710​WDIM​ + 37.87838​CIRCUM​ - 24.06528​FBEYE​ +
    13.36753​JAW

    Classification rule: Assign football player to ‘College’ if ​z1​ >


    ​ ​z​2​; else, assign football
    player to ‘HS’.

    3.
    a. Cross-validation yielded the table shown below:

    College HS Total

    College 7 3 10

    HS 3 7 10

    TOTAL 10 10 20

    3+3
    Misclassification Rate = 20 = 0.3

    ∴ ​The estimated misclassification rate using cross-validation is ​0.3 ​or​ 30%​.


    b. H​o​: The LDF has no discriminating ability.
    H​a​: The LDF has a discriminating ability.

    Test Procedure: F-test

    Decision Rule: Reject H​o ​if ​Fc​ ​ > F​0.05(4, 15)​; Otherwise, fail to reject H​o​.

    Computation:

    Let y​1​ = HS and y​2​ = College:


    Variable y​1 y​2 a​T Ӯ​1 Ӯ​2 Ӯ​1​-Ӯ​2 a​T​*(Ӯ​1​-Ӯ​2​)
    (y​1​-y​2​)

    WDIM 20.7271 21.15704 -0.42994 15.05 15.39 -0.34 0.1461796

    CIRCUM 37.87838 37.85851 0.01987 55.85 58.19 0.36 0.0071532

    FBEYE -24.06528 -24.60763 0.54235 20.37 20 0.37 0.2006695

    JAW 13.36753 15.75632 -2.38879 11.71 13.2 -1.49 3.5592971

    D​2 3.9132994

    F​c​ = ​(10 + 10 - 4 - 1) * (10)(10)​ * 3.9132994


    4*(10+10)*(10+10-2)
    F​c​ = 1​ 500​ * 3.9132994
    1440
    F​c​ = 1.0417 * 3.9132994
    F​c​ = ​4.0763535418 F​0.05(4, 15) =
    ​ 3.0556

    ​ 3.0556, we ​reject H​o​.


    Decision: Since F​c​ = 4.0764 > F​0.05(4, 15) =
    Conclusion: At 5% level of significance, there is sufficient evidence to say that
    the LDF has a discriminating ability​.
    4.

    Given ​WDIM​ = 15.7,​ CIRCUM​ = 59.1, ​FBEYE​ = 20.8, and ​JAW​ = 13.2:

    With respect to ​z1​ :​

    z​1 = ​ -​ 1122 + 21.15704​WDIM​ + 37.85851​CIRCUM​ - 24.60763​FBEYE​ + 15.75632​JAW


    z​1 =
    ​ - ​ 1122 + 21.15704​*(​15.7) + 37.85851​*(​59.1) - 24.60763​*(2​ 0.8) + 15.75632​*(​13.2)
    z​1 = ​ - ​ 1122 + 332.165528 + 2237.437941 - 511.838704 + 207.983424
    z​1 =​ 1 ​ 143.748189

    With respect to ​z2​ :​

    z​2 = ​ -​ 1098 + 20.72710​WDIM​ + 37.87838​CIRCUM​ - 24.06528​FBEYE​ + 13.36753​JAW


    z​2 =
    ​ - ​ 1098 + 20.72710​*(​15.7) + 37.87838​*(​59.1) - 24.06528​*(2​ 0.8) + 13.36753​*(​13.2)
    z​2 = ​ - ​ 1098 + 325.41547 + 2238.612258 - 500.557824 + 176.451396
    z​2 =​ 1 ​ 141.9213

    Conclusion: Since ​z​1 ​= 1​ 143.748189 > ​z2​ =


    ​ ​1141.9213, the group of a player with a head width at
    the widest dimension of 15.7, head circumference of 59.1, front-to-back measurement at the eye
    level of 20.8, and jaw width of 13.2 is ​COLLEGE​.

    You might also like