Matrix (version 1.2-8)

nearPD: Nearest Positive Definite Matrix

Description

Compute the nearest positive definite matrix to an approximate one, typically a correlation or variance-covariance matrix.

Usage

nearPD(x, corr = FALSE, keepDiag = FALSE, do2eigen = TRUE, doSym = FALSE, doDykstra = TRUE, only.values = FALSE, ensureSymmetry = !isSymmetric(x), eig.tol = 1e-06, conv.tol = 1e-07, posd.tol = 1e-08, maxit = 100, conv.norm.type = "I", trace = FALSE)

Arguments

x

numeric $n * n$ approximately positive definite matrix, typically an approximation to a correlation or covariance matrix. If x is not symmetric (and ensureSymmetry is not false), symmpart(x) is used.

corr

logical indicating if the matrix should be a correlation matrix.

keepDiag

logical, generalizing corr: if TRUE, the resulting matrix should have the same diagonal (diag(x)) as the input matrix.

do2eigen

logical indicating if a posdefify() eigen step should be applied to the result of the Higham algorithm.

doSym

logical indicating if X <- (X + t(X))/2 should be done, after X <- tcrossprod(Qd, Q); some doubt if this is necessary.

doDykstra

logical indicating if Dykstra's correction should be used; true by default. If false, the algorithm is basically the direct fixpoint iteration $Y(k) = P_U(P_S(Y(k-1)))$.

only.values

logical; if TRUE, the result is just the vector of eigen values of the approximating matrix.

ensureSymmetry

logical; by default, symmpart(x) is used whenever isSymmetric(x) is not true. The user can explicitly set this to TRUE or FALSE, saving the symmetry test. Beware however that setting it FALSE for an asymmetric input x, is typically nonsense!

eig.tol

defines relative positiveness of eigenvalues compared to largest one, $\lambda_1$. Eigen values $\lambda_k$ are treated as if zero when $\lambda_k / \lambda_1 \le eig.tol$.

conv.tol

convergence tolerance for Higham algorithm.

posd.tol

tolerance for enforcing positive definiteness (in the final posdefify step when do2eigen is TRUE).

maxit

maximum number of iterations allowed.

conv.norm.type

convergence norm type (norm(*, type)) used for Higham algorithm. The default is "I" (infinity), for reasons of speed (and back compatibility); using "F" is more in line with Higham's proposal.

trace

logical or integer specifying if convergence monitoring should be traced.

Value

If only.values = TRUE, a numeric vector of eigen values of the approximating matrix; Otherwise, as by default, an S3 object of class "nearPD", basically a list with components

Details

This implements the algorithm of Higham (2002), and then (if do2eigen is true) forces positive definiteness using code from posdefify. The algorithm of Knol DL and ten Berge (1989) (not implemented here) is more general in (1) that it allows constraints to fix some rows (and columns) of the matrix and (2) to force the smallest eigenvalue to have a certain value.

Note that setting corr = TRUE just sets diag(.) <- 1 within the algorithm.

Higham (2002) uses Dykstra's correction, but the version by Jens Oehlschlaegel did not use it (accidentally), and has still lead to good results; this simplification, now only via doDykstra = FALSE, was active in nearPD() upto Matrix version 0.999375-40.

References

Cheng, Sheung Hun and Higham, Nick (1998) A Modified Cholesky Algorithm Based on a Symmetric Indefinite Factorization; SIAM J. Matrix Anal.\ Appl., 19, 1097--1110.

Knol DL, ten Berge JMF (1989) Least-squares approximation of an improper correlation matrix by a proper one. Psychometrika 54, 53--61.

Higham, Nick (2002) Computing the nearest correlation matrix - a problem from finance; IMA Journal of Numerical Analysis 22, 329--343.

See Also

A first version of this (with non-optional corr=TRUE) has been available as nearcor(); and more simple versions with a similar purpose posdefify(), both from package sfsmisc.

Examples

 ## Higham(2002), p.334f - simple example
 A <- matrix(1, 3,3); A[1,3] <- A[3,1] <- 0
 n.A <- nearPD(A, corr=TRUE, do2eigen=FALSE)
 n.A[c("mat", "normF")]
 stopifnot(all.equal(n.A$mat[1,2], 0.760689917),
	   all.equal(n.A$normF, 0.52779033, tolerance=1e-9) )

 set.seed(27)
 m <- matrix(round(rnorm(25),2), 5, 5)
 m <- m + t(m)
 diag(m) <- pmax(0, diag(m)) + 1
 (m <- round(cov2cor(m), 2))

 str(near.m <- nearPD(m, trace = TRUE))
 round(near.m$mat, 2)
 norm(m - near.m$mat) # 1.102 / 1.08

 if(require("sfsmisc")) {
    m2 <- posdefify(m) # a simpler approach
    norm(m - m2)  # 1.185, i.e., slightly "less near"
 }

 round(nearPD(m, only.values=TRUE), 9)

## A longer example, extended from Jens' original,
## showing the effects of some of the options:

pr <- Matrix(c(1,     0.477, 0.644, 0.478, 0.651, 0.826,
               0.477, 1,     0.516, 0.233, 0.682, 0.75,
               0.644, 0.516, 1,     0.599, 0.581, 0.742,
               0.478, 0.233, 0.599, 1,     0.741, 0.8,
               0.651, 0.682, 0.581, 0.741, 1,     0.798,
               0.826, 0.75,  0.742, 0.8,   0.798, 1),
             nrow = 6, ncol = 6)

nc.  <- nearPD(pr, conv.tol = 1e-7) # default
nc.$iterations  # 2
nc.1 <- nearPD(pr, conv.tol = 1e-7, corr = TRUE)
nc.1$iterations # 11 / 12 (!)
ncr   <- nearPD(pr, conv.tol = 1e-15)
str(ncr)# still 2 iterations
ncr.1 <- nearPD(pr, conv.tol = 1e-15, corr = TRUE)
ncr.1 $ iterations # 27 / 30 !

ncF <- nearPD(pr, conv.tol = 1e-15, conv.norm = "F")
stopifnot(all.equal(ncr, ncF))# norm type does not matter at all in this example

## But indeed, the 'corr = TRUE' constraint did ensure a better solution;
## cov2cor() does not just fix it up equivalently :
norm(pr - cov2cor(ncr$mat)) # = 0.09994
norm(pr -       ncr.1$mat)  # = 0.08746 / 0.08805

### 3) a real data example from a 'systemfit' model (3 eq.):
(load(system.file("external", "symW.rda", package="Matrix"))) # "symW"
dim(symW) #  24 x 24
class(symW)# "dsCMatrix": sparse symmetric
if(dev.interactive())  image(symW)
EV <- eigen(symW, only=TRUE)$values
summary(EV) ## looking more closely {EV sorted decreasingly}:
tail(EV)# all 6 are negative
EV2 <- eigen(sWpos <- nearPD(symW)$mat, only=TRUE)$values
stopifnot(EV2 > 0)
if(require("sfsmisc")) {
       plot(pmax(1e-3,EV), EV2, type="o", log="xy", xaxt="n",yaxt="n")
       eaxis(1); eaxis(2)
} else plot(pmax(1e-3,EV), EV2, type="o", log="xy")
abline(0,1, col="red3",lty=2)