Information Technology Reference
In-Depth Information
Figure 9. Sample input and output of multi-layer relationship pattern
0
1
2 3 4
5 6 7 8
9 10
0
90
80
1
2
3
4
5
6
7
8
70
9
10
0
1
2 3 4
5 6 7 8
9 10
0
70
75
1
2
3
4
5
6
7
8
85
9
10
threshold. The advantage of doing so is that by using different value of confidence threshold in each
layer, decision makers can adjust the confidence thresholds in each layer in order to suit their deci-
sion needs. For decisions that need to have a very strong confidence, the confidence threshold can be
increased, while for decisions that need to have moderate level of confidence, the confidence threshold
could be adjusted down.
Once the confidence threshold for each layer of matrices is obtained, the relationship between static
nodes can be compared. Within the same layer, the relationships between the static nodes are defined
as the addition of both confidences divided by two. Relationship between static nodes within the same
matrix layer: Relationship = [
Conidence
(0,0) +
Conidence
(8,8)] /
2
.
Figure 9 shows the result of multi-layer matrix based mining method with a matrix size of
10*10*7
matrix spaces. The upper matrix is fashion domain and the bottom matrix is leisure domain. It measures
the confidence between two domains that is fashion and leisure. In the upper matrix, three relationships
can be found under the fashion domain. They are: (0,0 - 8,0) = 85, (0,0 - 1,8) = 80, (8,0 - 1,8) = 75. In
the lower matrix, three relationships can be found under the leisure domain. They are: (8,0 - 1,8) = 80,
Search WWH ::
Custom Search