Information Technology Reference
In-Depth Information
ter represented in a discrete form by using
x
axis and
y
axis. The benefit of this enhances the pinpoint
of location in the mobile environment (Goh & Taniar, 2004b), and it will be used to find the hot spots
where a mobile user frequents.
Matrix (Ashcraft & Liu, 1998) is a concept from discrete mathematics. It is a list of items, such as
numbers, arranged in a pre-defined sequence which has the width (
m
) and length (
n
) specified (Goh &
Taniar, 2004b). The number of items available for storing into a matrix is equal to the product of m and
n. For example, if
m
=
10
, and
n
=
10
, the number of storage space available in this
10*10
matrix is
100
.
A matrix (Ashcraft & Liu, 1998) can be categorized into dense matrix or sparse matrix. The dense and
sparse categorisation is done by looking at how dense the dataset appears in a matrix. If the matrix has
meaningful data filled more than 50% of the matrix, it is referred to as dense matrix (Ashcraft & Liu,
1998), and if the matrix has meaningful data filled at less than 50%, it is considered as sparse matrix
(Ashcraft & Liu, 1998).
The issue of dense or sparse matrix (Ashcraft & Liu, 1998) is of little concern in this proposed
method, as the matrix is used as a method for mapping the physical location map to itself, and serves
as a physical location divider and numbering system provider. Therefore, in this proposed method, the
matrix (Ashcraft & Liu, 1998) can either be dense or sparse. However, it would be a sparse matrix most
of the time, as items of interest to the data miner and related to the problem to be solved should be at
a distance from each others, otherwise, if they are very near, they should be considered as one single
unit rather than two separate units.
Pro Posed Method: Matri X Pattern
A matrix is a storage component which contains multiple storage areas across the horizontal and verti-
cal paths in a two dimensional plane (Ashcraft & Liu, 1998). The position of each storage location in
the matrix is fixed, and all location must have a value stored. The size of the matrix that is the length of
the horizontal and vertical plane can be adjusted depending on the nature of problem to be solved. It is
important to note that matrix consists of multiple storage areas and the purpose of doing so will greatly
enhance the efficiency of problem solving.
In this paper, we propose a method for using matrix to solve the problem of mobile user data mining.
The problem of mobile user data mining is to address the challenges of using source data obtained from
mobile users in order to find out useful patterns. These patterns can then be translated into knowledge,
whenever the format of the pattern is disclosed. In order to meet the challenge, matrix pattern is pro-
posed.
Matrix pattern is represented as a single matrix or multiple matrices in which each storage area in
the matrix is representing a single static node. An example of static node is the wireless network access
point, which provides the bandwidth to the mobile users. In order to provide the pattern of the behaviour
of mobile users, the relationships in between each storage areas of the matrix is defined. The relation-
ship between two storage areas in the matrix is defined as:
X
→
Y
= {
0%
→
100%
}. The relationship is
directional, which translates to
X
→
Y
≠
Y
→
X
. Each relationship comes with a magnitude of deepness
of the relationship, which is represented as a percentage with lowest value of
0%
and highest value of
100%
.
Each of the storage area in the matrix represents a single static node in the mobile environment. Static
nodes are widely available, and their positions are often accurately documented. Due to the fact that it
Search WWH ::
Custom Search