dc.description.abstract |
Multidimensional
scaling
is a technique
used
to picture,
visualize
or locate
objects
in a
map
or
Shepard
diagram,
Buja
et ai,
(2004).
In
this
dissertation
multidimensional
i
I
.
scaling
method
has
been
applied
to visualize
approximate
Iodations
of eleven
sampled
i
local
governments
in
Kebbi
State.
The
symmetric
distance
matrix
'D'
has
been
processed
from
an
atlas
of Nigeria
by
Eunice,
(1979).
The
data
has
been
transformed
into
inner-product
matrix
B by
double
centering,
using
Mardis
et al
(1979),
Su
and
Chang
(2000),
Hardle
and
Simax
(2003)
methodology
to arrive
at the
required
results.
Matrix
B has
been
analyzed
using
MATLAB
Software
to:
obtain
eigenvalues
and
eigenvectors.
From
the
results
of
analysis,
the
first
2 eigenvalues
are
large
and
positive,
which
signifies
goodness
of fit (strain
measure).
It also
indicates
that
it is a 2-
I
dimensional
plot
and
'R'
package
has
been
used
in plotting
the
approximate
locations. |
en_US |