Mining Matrix Pattern from Mobile Users - Distributed Artificial Intelligence, Agent Technology, and Collaborative Applications - page 249

to mining by using mobile nodes because of the extra processing power required to track the mobile
nodes and gather data from them. Static nodes, being well equipped with processing power and storage
resources keep track of mobile nodes at all times by providing resources to them. These static nodes
are in the best position to find out knowledge from mobile users as it stays in touch with the mobile
nodes at all times.
However, the benefit of using the matrix pattern to be cost efficient comes with a price depending
on interpretation. The matrix pattern, a location-dependent data mining method, generates knowledge
that represents the pattern among static nodes. Therefore, it will provide a physical location based
kind of knowledge. To enhance this, the static nodes in the matrix pattern are named with their logical
themes. These logical themes will then be able to be used to judge the relationship among the static
nodes better.
The performance evaluation is done on a Pentium IV machine with 384MB of RAM and 10GB of hard
disk space. The datasets used in the performance evaluation are derived from synthetic data generator.
Three datasets were used, namely equal , double and triple . Equal , double and triple have different sizes

Figure 16. Comparison among single, double, and triple y coordinates

3000

2500

2000

1500

1000

500

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

No of XY Coordinates

Single Double Triple

3000

2500

2000

1500

1000

500

0

Triple Equal Double

Types of Data

X=10 X=20 X=30