In | The n by n identity matrix |
1n | The n by 1 vector of unit elements |
O | A matrix of appropriate order with all zero entries |
Λi | The ith largest eigenvalue of the matrix under consideration |
|A| | The determinant of the square matrix A |
tr(A) | The trace of the square matrix A |
A−1 | The inverse of the matrix A |
A1/2 | The symmetric square root of the matrix A |
A− | A generalized inverse of the matrix A |
E(y) | Expected value of a random variable or vector y |
v(y), var(y) | Variance of a random variable y |
cov(x, y) | Covariance of random variable (vector) x with random variable (or vector) y |
D(y) | The variance covariance or the dispersion matrix of y |
Np (μ, Σ) | A p-dimensional normal distribution with mean μ and the variance covariance matrix Σ |
Wp (ƒ, Σ) | A p-(matrix) variate Wishart distribution with ƒ degrees of freedom and parameter Σ (that is, with expected value ƒ Σ) |
ε | Error vector |
ϵ | Error matrix |
Y | n by p matrix of data on dependent variables |
X | Regression/Design matrix in the linear model |
β | Regression/Design parameter vector |
B | Regression/Design parameter matrix |
Σ | (usually) The Dispersion matrix of errors |
df | Degrees of freedom |
SS&CP Matrix | Matrix of the sums of squares and crossproducts |
E | Error SS&CP matrix |
H | Hypothesis SS&CP matrix |
The sample mean vector | |
S | Sample dispersion matrix (with d f as denominator) |
Sn | Sample dispersion matrix (with sample size as denominator) |
P | The Projection or Hat matrix |
T2 | Hotelling's T2 |
Λ | Wilks' Lambda |
β1,p | Coefficient of multivariate skewness |
β1,p | Coefficient of multivariate kurtosis |
⊗ | Kronecker product |
AIC | Akaike's information criterion |
BIC | Swartz's Bayesian information criterion |
3.143.4.181