1
<?xml version="1.0" encoding="ISO-8859-1" standalone="no"?>
2
<!DOCTYPE MAN SYSTEM "../../manrev.dtd">
4
<LANGUAGE>eng</LANGUAGE>
6
<TYPE>Scilab Function</TYPE>
7
<DATE>April 1993</DATE>
8
<SHORT_DESCRIPTION name="hank"> covariance to hankel matrix</SHORT_DESCRIPTION>
10
<CALLING_SEQUENCE_ITEM>[hk]=hank(m,n,cov) </CALLING_SEQUENCE_ITEM>
15
<PARAM_NAME>m</PARAM_NAME>
17
<SP>: number of bloc-rows</SP>
21
<PARAM_NAME>n</PARAM_NAME>
23
<SP>: number of bloc-columns</SP>
27
<PARAM_NAME>cov</PARAM_NAME>
29
<SP>: sequence of covariances; it must be given as :[R0 R1 R2...Rk]</SP>
33
<PARAM_NAME>hk</PARAM_NAME>
35
<SP>: computed hankel matrix</SP>
42
this function builds the hankel matrix of size <VERB>(m*d,n*d)</VERB>
43
from the covariance sequence of a vector process</P>
47
//Example of how to use the hank macro for
48
//building a Hankel matrix from multidimensional
49
//data (covariance or Markov parameters e.g.)
51
//This is used e.g. in the solution of normal equations
52
//by classical identification methods (Instrumental Variables e.g.)
54
//1)let's generate the multidimensional data under the form :
55
// C=[c_0 c_1 c_2 .... c_n]
56
//where each bloc c_k is a d-dimensional matrix (e.g. the k-th correlation
57
//of a d-dimensional stochastic process X(t) [c_k = E(X(t) X'(t+k)], '
58
//being the transposition in scilab)
60
//we take here d=2 and n=64
64
//generate the hankel matrix H (with 4 bloc-rows and 5 bloc-columns)
76
<AUTHOR>G. Le Vey</AUTHOR>