STAT 3690 Lecture 15
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
STAT 3690 Lecture 15
2022
Testing for equality of covariance matrices (J&W Sec. 6.6)
● Model: m independent samples, where
- x11 , . . . , x1n1 MVNp (u1 , 51 )
-
- xm1 , . . . , xmn. MVNp (um , 5m )
● Hypotheses H0 : 51 = . . . = 5m v.s. H1 : otherwise
● MLE of (u1 , . . . , um , 51 , . . . , 5m )
- Under H0
* i = i = n xij
* i = ( i ni )_ 1 SSPw = ( i ni )_ 1 ij (xij _ i )(xij _ i )T for all i
- No restriction on 5i
* i = i = n xij
* i = ni(_)1 (ni _ 1)Si = n (xij _ i )(xij _ i )T
● Likelihood ratio
λ = i ┌ ┐nm /2
- Spool = ( i ni _ m)_ 1 SSPw
M = _2 ln
i
╱ \(nm _ 1)/2
- Under H0
(1 _ u)M s χ2 (p(p + 1)(m _ 1)/2)
* u = ( i (ni _ 1)_ 1 _ ( i ni _ m)_ 1 }(6(p + 1)(m _ 1)}_ 1 (2p2 + 3p _ 1)
● Rejection region at level α
…x11 , . . . , x1n1 , x21 , . . . , xmn. : (1 _ u)M > χ1(2) _α,p(p+1)(m _ 1)/2入
● p-value
1 _ Fχ1(2) − a,p(p+1)(. − 1)/2 ((1 _ u)M }
● Exercise: factors in producing plastic film (continued)
- Check the equality of covariance matrices for RATE="Low" and RATE="High"
install.packages( 'heplots ')
options(digits = 4)
tear <- c (
6.5 , 6.2 , 5.8 , 6.5 , 6.5 , 6.9 , 7.2 , 6.9 , 6.1 , 6.3 ,
6.7 , 6.6 , 7.2 , 7.1 , 6.8 , 7.1 , 7.0 , 7.2 , 7.5 , 7.6
)
gloss <- c (
9.5 , 9.9 , 9.6 , 9.6 , 9.2 , 9.1 , 10.0 , 9.9 , 9.5 , 9.4 ,
9.1 , 9.3 , 8.3 , 8.4 , 8.5 , 9.2 , 8.8 , 9.7 , 10.1 , 9.2
)
opacity <- c (
4.4 , 6.4 , 3.0 , 4.1 , 0.8 , 5.7 , 2.0 , 3.9 , 1.9 , 5.7 ,
2.8 , 4.1 , 3.8 , 1.6 , 3.4 , 8.4 , 5.2 , 6.9 , 2.7 , 1.9
)
(X <- cbind(tear, gloss, opacity))
(rate <- factor(gl(2 ,10 ,length=nrow (X)), labels=c ( "Low" , "High")))
(additive <- factor(gl(2 ,5 ,length=nrow (X)), labels=c ( "Low" , "High")))
result = heplots ::boxM(lm(X~rate))
result$statistic
result$p.value
● Report: Testing hypotheses H0 : the covariance matrix does not vary with the level of RATE v.s. H1 : otherwise, we carried on the Box’s M test and obtained 4.017 as the value of test statistic. The corresponding p-value was 0.6743. So, at the .05 level, there was no strong statistical evidence against H0 , i.e., we did not reject H0 and believed that the covariance matrix does not vary with the level of RATE.
2022-03-18