simulAR1 {G1DBN} | R Documentation |
This function generates multivariate time series according to the following first order Auto-Regressive process,
X(t)= A X(t-1) + B + e(t),
where matrix A has size p x p and arrays X(t), B and e(t) have length p. e(t) follows a zero-centered multivariate gaussian distribution whose variance matrix S is diagonal. First, matrix A, array B and diagonal of S are randomly generated. Each diagonal term S[i,i] is uniformly generated from U([minSig,maxSig]). The elements of matrix A and array B are uniformly generated from U([-maxA,-minA],[minA,maxA]) and U([minB,maxB]) respectively. Second, the time series data are generated according the so defined AR(1) model.
out<-simulAR1(p,n,edgeProp,minA,maxA,minB,maxB,minSig,maxSig)
p |
the desired dimension of the multivariate time series. |
n |
the desired length of the time serie. |
edgeProp |
the desired proportion of non zero coefficient in the AR transition matrix. |
minA |
the minimum value for matrix A elements generation. |
maxA |
the maximum value for matrix A elements generation. |
minB |
the minimum value for matrix B elements generation. |
maxB |
the maximum value for matrix B elements generation. |
minSig |
the minimum value for the diagonal of covariance matrix S generation. |
maxSig |
the maximum value for the diagonal of covariance matrix S generation. |
A list with out$data a matrix, with n rows (=length) and p columns (=dimension), containing the generated time series, out$A the AR generated matrix A (p x p), out$B the AR generated vector B (p), out$sig the generated diagonal (p) of covariance matrix S.
Lebre Sophie (http://stat.genopole.cnrs.fr/~slebre).
#generate AR(1) time series AR<-simulAR1(p=10,n=50,edgeProp=0.02,minA=0.5,maxA=1.5,minB=0,maxB=1,minSig=0.1,maxSig=0.8)