← Back to branch summary

~ubuntu-branches/ubuntu/hoary/scilab/hoary

« back to all changes in this revision

Viewing changes to man/fr/signal/rpem.xml

  • Committer: Bazaar Package Importer
  • Author(s): Torsten Werner
  • Date: 2005-01-09 22:58:21 UTC
  • mfrom: (1.1.1 upstream)
  • Revision ID: james.westby@ubuntu.com-20050109225821-473xr8vhgugxxx5j
Tags: 3.0-12
http://bugs.debian.org/255869
changed configure.in to build scilab's own malloc.o, closes: #255869

Show diffs side-by-side

added added

removed removed

Lines of Context:
man/fr/signal/rpem.xml
 
1
<?xml version="1.0" encoding="ISO-8859-1" standalone="no"?>
 
2
<!DOCTYPE MAN SYSTEM "../../manrev.dtd">
 
3
<MAN>
 
4
  <LANGUAGE>eng</LANGUAGE>
 
5
  <TITLE>rpem</TITLE>
 
6
  <TYPE>Scilab Function</TYPE>
 
7
  <DATE>April 1993</DATE>
 
8
  <SHORT_DESCRIPTION name="rpem"> RPEM estimation</SHORT_DESCRIPTION>
 
9
  <CALLING_SEQUENCE>
 
10
    <CALLING_SEQUENCE_ITEM>[w1,[v]]=rpem(w0,u0,y0,[lambda,[k,[c]]])  </CALLING_SEQUENCE_ITEM>
 
11
  </CALLING_SEQUENCE>
 
12
  <PARAM>
 
13
    <PARAM_INDENT>
 
14
      <PARAM_ITEM>
 
15
        <PARAM_NAME>a,b,c</PARAM_NAME>
 
16
        <PARAM_DESCRIPTION>
 
17
          <SP>: <VERB>a=[a(1),...,a(n)], b=[b(1),...,b(n)], c=[c(1),...,c(n)]</VERB></SP>
 
18
        </PARAM_DESCRIPTION>
 
19
      </PARAM_ITEM>
 
20
      <PARAM_ITEM>
 
21
        <PARAM_NAME>w0</PARAM_NAME>
 
22
        <PARAM_DESCRIPTION>
 
23
          <SP>: <VERB>list(theta,p,phi,psi,l)</VERB> where:</SP>
 
24
          <PARAM_INDENT>
 
25
            <PARAM_ITEM>
 
26
              <PARAM_NAME>theta</PARAM_NAME>
 
27
              <PARAM_DESCRIPTION>
 
28
                <SP>: [a,b,c] is a real vector of order <VERB>3*n</VERB></SP>
 
29
              </PARAM_DESCRIPTION>
 
30
            </PARAM_ITEM>
 
31
            <PARAM_ITEM>
 
32
              <PARAM_NAME>p</PARAM_NAME>
 
33
              <PARAM_DESCRIPTION>
 
34
                <SP>: (3*n x 3*n) real matrix.</SP>
 
35
              </PARAM_DESCRIPTION>
 
36
            </PARAM_ITEM>
 
37
            <PARAM_ITEM>
 
38
              <PARAM_NAME>phi,psi,l</PARAM_NAME>
 
39
              <PARAM_DESCRIPTION>
 
40
                <SP>: real vector of dimension <VERB>3*n</VERB></SP>
 
41
              </PARAM_DESCRIPTION>
 
42
            </PARAM_ITEM>
 
43
          </PARAM_INDENT>
 
44
          <P>
 
45
    During the first call on can take:
 
46
  </P>
 
47
          <VERBATIM>
 
48
<![CDATA[
 
49
theta=phi=psi=l=0*ones(1,3*n). p=eye(3*n,3*n)
 
50
   ]]>
 
51
          </VERBATIM>
 
52
        </PARAM_DESCRIPTION>
 
53
      </PARAM_ITEM>
 
54
      <PARAM_ITEM>
 
55
        <PARAM_NAME>u0</PARAM_NAME>
 
56
        <PARAM_DESCRIPTION>
 
57
          <SP>: real vector of inputs (arbitrary size) (if no input take <VERB>u0=[ ]</VERB>).</SP>
 
58
        </PARAM_DESCRIPTION>
 
59
      </PARAM_ITEM>
 
60
      <PARAM_ITEM>
 
61
        <PARAM_NAME>y0</PARAM_NAME>
 
62
        <PARAM_DESCRIPTION>
 
63
          <SP>: vector of outputs (same dimension as <VERB>u0</VERB> if <VERB>u0</VERB> is not empty). (<VERB>y0(1)</VERB> is not used by rpem).</SP>
 
64
          <P>
 
65
    If the time domain is <VERB>(t0,t0+k-1)</VERB> the <VERB>u0</VERB> vector contains the inputs 
 
66
  </P>
 
67
          <P><VERB>u(t0),u(t0+1),..,u(t0+k-1)</VERB> and <VERB>y0</VERB> the outputs 
 
68
  </P>
 
69
          <P>
 
70
            <VERB>y(t0),y(t0+1),..,y(t0+k-1)</VERB>
 
71
          </P>
 
72
        </PARAM_DESCRIPTION>
 
73
      </PARAM_ITEM>
 
74
    </PARAM_INDENT>
 
75
  </PARAM>
 
76
  <DESCRIPTION>
 
77
    <P>
 
78
    Recursive estimation of parameters in an ARMAX model.
 
79
    Uses Ljung-Soderstrom recursive prediction error method.
 
80
    Model considered is the following:
 
81
  </P>
 
82
    <VERBATIM>
 
83
<![CDATA[
 
84
y(t)+a(1)*y(t-1)+...+a(n)*y(t-n)=
 
85
b(1)*u(t-1)+...+b(n)*u(t-n)+e(t)+c(1)*e(t-1)+...+c(n)*e(t-n)
 
86
   ]]>
 
87
    </VERBATIM>
 
88
    <P>
 
89
  </P>
 
90
    <P>
 
91
    The effect of this command is to update the estimation of
 
92
    unknown parameter <VERB>theta=[a,b,c]</VERB> with
 
93
  </P>
 
94
    <P><VERB>a=[a(1),...,a(n)], b=[b(1),...,b(n)], c=[c(1),...,c(n)]</VERB>.
 
95
  </P>
 
96
  </DESCRIPTION>
 
97
  <SECTION label="Optional parameters">
 
98
    <ITEMIZE>
 
99
      <ITEM label="lambda">
 
100
        <SP>: optional parameter (forgetting constant) choosed close to 1 as convergence occur:</SP>
 
101
        <P><VERB>lambda=[lambda0,alfa,beta]</VERB> evolves according to :</P>
 
102
        <VERBATIM>
 
103
<![CDATA[
 
104
lambda(t)=alfa*lambda(t-1)+beta 
 
105
   ]]>
 
106
        </VERBATIM>
 
107
        <P>
 
108
    with <VERB>lambda(0)=lambda0</VERB></P>
 
109
        <P><VERB>k</VERB> : contraction factor to be chosen close to 1 as convergence occurs.</P>
 
110
        <P><VERB>k=[k0,mu,nu]</VERB> evolves according to:</P>
 
111
        <VERBATIM>
 
112
<![CDATA[
 
113
k(t)=mu*k(t-1)+nu 
 
114
   ]]>
 
115
        </VERBATIM>
 
116
        <P>
 
117
    with <VERB>k(0)=k0</VERB>.</P>
 
118
        <P><VERB>c</VERB> : large parameter.(<VERB>c=1000</VERB> is the default value).</P>
 
119
      </ITEM>
 
120
    </ITEMIZE>
 
121
  </SECTION>
 
122
  <SECTION label="Output parameters:">
 
123
    <P><VERB>w1</VERB>: update for <VERB>w0</VERB>.</P>
 
124
    <P><VERB>v</VERB>: sum of squared prediction errors on <VERB>u0, y0</VERB>.(optional).</P>
 
125
    <P>
 
126
    In particular <VERB>w1(1)</VERB> is the new estimate of <VERB>theta</VERB>. If a new sample <VERB>u1, y1</VERB> is available the update is obtained by:</P>
 
127
    <P><VERB>[w2,[v]]=rpem(w1,u1,y1,[lambda,[k,[c]]])</VERB>. Arbitrary large series can thus be treated.</P>
 
128
  </SECTION>
 
129
</MAN>