Integer type: int32 int64 nag_int show int32 show int32 show int64 show int64 show nag_int show nag_int

PDF version (NAG web site, 64-bit version, 64-bit version)

Chapter Contents

Chapter Introduction

NAG Toolbox

NAG Toolbox: nag_mv_distance_mat (g03ea)

▸▿ Contents

1 Purpose

2 Syntax

3 Description

4 References

▸▿ 5 Parameters

5.1 Compulsory Input Parameters

5.2 Optional Input Parameters

5.3 Output Parameters

6 Error Indicators and Warnings

7 Accuracy

8 Further Comments

9 Example

Purpose

nag_mv_distance_mat (g03ea) computes a distance (dissimilarity) matrix.

Syntax

[s, d, ifail] = g03ea(update, dist, scal, x, isx, s, d, 'n', n, 'm', m)

[s, d, ifail] = nag_mv_distance_mat(update, dist, scal, x, isx, s, d, 'n', n, 'm', m)

Note: the interface to this routine has changed since earlier releases of the toolbox:

At Mark 22:

n was made optional

Description

Given

n

objects, a distance or dissimilarity matrix is a symmetric matrix with zero diagonal elements such that the

i j

th element represents how far apart or how dissimilar the

i

th and

j

th objects are.

Let

X

be an

n

p

data matrix of observations of

p

variables on

n

objects, then the distance between object

j

and object

k

d_{j k}

, can be defined as:

d_{j k} = {\{\sum_{i = 1}^{p} D (x_{j i} / s_{i}, x_{k i} / s_{i})\}}^{α},

where

x_{j i}

and

x_{k i}

are the

j i

th and

k i

th elements of

X

s_{i}

is a standardization for the

i

th variable and

D (u, v)

is a suitable function. Three functions are provided in nag_mv_distance_mat (g03ea).

(a)	Euclidean distance: $D (u, v) = {(u - v)}^{2}$ and $α = \frac{1}{2}$ .
(b)	Euclidean squared distance: $D (u, v) = {(u - v)}^{2}$ and $α = 1$ .
(c)	Absolute distance (city block metric): $D (u, v) = \|u - v\|$ and $α = 1$ .

Three standardizations are available.

(a)	Standard deviation: $s_{i} = \sqrt{\sum_{j = 1}^{n} {(x_{j i} - \bar{x})}^{2} / (n - 1)}$
(b)	Range: $s_{i} = \max (x_{1 i}, x_{2 i}, \dots, x_{n i}) - \min (x_{1 i}, x_{2 i}, \dots, x_{n i})$
(c)	User-supplied values of $s_{i}$ .

In addition to the above distances there are a large number of other dissimilarity measures, particularly for dichotomous variables (see Krzanowski (1990) and Everitt (1974)). For the dichotomous case these measures are simple to compute and can, if suitable scaling is used, be combined with the distances computed by nag_mv_distance_mat (g03ea) using the updating option.

Dissimilarity measures for variables can be based on the correlation coefficient for continuous variables and contingency table statistics for dichotomous data, see chapters G02 and G11 respectively.

nag_mv_distance_mat (g03ea) returns the strictly lower triangle of the distance matrix.

References

Everitt B S (1974) Cluster Analysis Heinemann

Krzanowski W J (1990) Principles of Multivariate Analysis Oxford University Press

Parameters

Compulsory Input Parameters

1: $update$ – string (length ≥ 1)

Indicates whether or not an existing matrix is to be updated.

$update ='U'$: The matrix $D$ is updated and distances are added to $D$ .
$update ='I'$: The matrix $D$ is initialized to zero before the distances are added to $D$ .

Constraint:

update ='U'

'I'

2: $dist$ – string (length ≥ 1)

Indicates which type of distances are computed.

$dist ='A'$: Absolute distances.
$dist ='E'$: Euclidean distances.
$dist ='S'$: Euclidean squared distances.

Constraint:

dist ='A'

'E'

'S'

3: $scal$ – string (length ≥ 1)

Indicates the standardization of the variables to be used.

$scal ='S'$: Standard deviation.
$scal ='R'$: Range.
$scal ='G'$: Standardizations given in array s.
$scal ='U'$: Unscaled.

Constraint:

scal ='S'

'R'

'G'

'U'

4: $x (ldx, m)$ – double array

ldx, the first dimension of the array, must satisfy the constraint

ldx \geq n

x (i, j)

must contain the value of the

j

th variable for the

i

th object, for

i = 1, 2, \dots, n

and

j = 1, 2, \dots, m

5: $isx (m)$ – int64int32nag_int array

isx (j)

indicates whether or not the

j

th variable in x is to be included in the distance computations.

isx (j) > 0

the

j

th variable is included, for

j = 1, 2, \dots, m

; otherwise it is not referenced.

Constraint:

isx (j) > 0

for at least one

j

, for

j = 1, 2, \dots, m

6: $s (m)$ – double array

scal ='G'

and

isx (j) > 0

then

s (j)

must contain the scaling for variable

j

, for

j = 1, 2, \dots, m

Constraint: if

scal ='G'

and

isx (j) > 0

s (j) > 0.0

, for

j = 1, 2, \dots, m

7: $d (n \times (n - 1) / 2)$ – double array

update ='U'

, d must contain the strictly lower triangle of the distance matrix

D

to be updated.

D

must be stored packed by rows, i.e.,

d ((i - 1) (i - 2) / 2 + j)

i > j

must contain

d_{i j}

update ='I'

, d need not be set.

Constraint: if

update ='U'

d (j) \geq 0.0

, for

j = 1, 2, \dots, n (n - 1) / 2

Optional Input Parameters

1: $n$ – int64int32nag_int scalar: Default: the first dimension of the array x.
$n$ , the number of observations.

Constraint: $n \geq 2$ .
2: $m$ – int64int32nag_int scalar: Default: the dimension of the arrays isx, s and the second dimension of the array x. (An error is raised if these dimensions are not equal.)
The total number of variables in array x.

Constraint: $m > 0$ .

Output Parameters

1: $s (m)$ – double array: If $scal ='S'$ and $isx (j) > 0$ then $s (j)$ contains the standard deviation of the variable in the $j$ th column of x.
If $scal ='R'$ and $isx (j) > 0$ , $s (j)$ contains the range of the variable in the $j$ th column of x.

If $scal ='U'$ and $isx (j) > 0$ , $s (j) = 1.0$ .

If $scal ='G'$ , s is unchanged.
2: $d (n \times (n - 1) / 2)$ – double array: The strictly lower triangle of the distance matrix $D$ stored packed by rows, i.e., $d_{i j}$ is contained in $d ((i - 1) (i - 2) / 2 + j)$ , $i > j$ .
3: $ifail$ – int64int32nag_int scalar: $ifail = 0$ unless the function detects an error (see Error Indicators and Warnings).

Error Indicators and Warnings

Errors or warnings detected by the function:

$ifail = 1$

On entry,	$n < 2$ ,
or	$ldx < n$ ,
or	$m \leq 0$ ,
or	$update \neq'I'$ or $'U'$ ,
or	$dist \neq'A'$ , $'E'$ or $'S'$ ,
or	$scal \neq'S'$ , $'R'$ , $'G'$ or $'U'$ .

$ifail = 2$

On entry,	$isx (j) \leq 0$ , for $j = 1, 2, \dots, m$ ,
or	$update ='U'$ and $d (j) < 0.0$ , for some $j = 1, 2, \dots, n (n - 1) / 2$ ,
or	$scal ='S'$ or $'R'$ and $x (i, j) = x (i + 1, j)$ for $i = 1, 2, \dots, n - 1$ , for some $j$ with $isx (i) > 0$ .
or	$s (j) \leq 0.0$ for some $j$ when $scal ='G'$ and $isx (j) > 0$ .

$ifail = - 99$: An unexpected error has been triggered by this routine. Please contact NAG.

$ifail = - 399$: Your licence key may have expired or may not have been installed correctly.

$ifail = - 999$: Dynamic memory allocation failed.

Accuracy

The computations are believed to be stable.

Further Comments

nag_mv_cluster_hier (g03ec) can be used to perform cluster analysis on the computed distance matrix.

Example

A data matrix of five observations and three variables is read in and a distance matrix is calculated from variables

2

and

3

using squared Euclidean distance with no scaling. This matrix is then printed.

Open in the MATLAB editor: g03ea_example

function g03ea_example


fprintf('g03ea example results\n\n');

x = [1, 1, 1;
     2, 1, 2;
     3, 6, 3;
     4, 8, 2;
     5, 8, 0];
[n,m]  = size(x);

isx    = ones(m,1,'int64');
isx(1) = int64(0);
s      = ones(m,1);
ld     = (n*(n-1))/2;
d      = zeros(ld,1);

% Compute the distance matrix
update = 'I';
dist = 'S';
scal = 'U';
[s, d, ifail] = g03ea( ...
		       update, dist, scal, x, isx, s, d);

fprintf(' Distance Matrix\n  ');
fprintf('   %5d', [1:n-1]);
for i = 2:n
  lj = (i-1)*(i-2)/2 + 1;
  uj = i*(i-1)/2;
  fprintf('\n%2d  ', i);
  fprintf('   %5.2f', d(lj:uj));
end
fprintf('\n');

g03ea example results

 Distance Matrix
         1       2       3       4
 2      1.00
 3     29.00   26.00
 4     50.00   49.00    5.00
 5     50.00   53.00   13.00    4.00

PDF version (NAG web site, 64-bit version, 64-bit version)

Chapter Contents

Chapter Introduction

NAG Toolbox