Information Technology Reference
In-Depth Information
Fig. 3.
Standard retrieved pattern from process (c) and (b)
Fig. 4.
Standard retrieved pattern from process (d) and (b)
in-between each correspondence set, we figure out the similar pattern in pair of
Process (a) and (b) is (
).
Note for calculating the precision rate, we only considers the nodes in the cor-
respondence sets. The rest figures are analog.
{a
1
,a
2
,a
3
,a
4
}
,
{b
1
, {E,F,G,H} ,b
2
, {J, K, L} ,b
3
,b}
Tabl e 1.
Gold Standard of Retrieved Patterns
Process model pair
Similar patterns
(a) and (b)
{a
1
,a
2
,a
3
,a
4
} and {b
1
, {E,F,G,H} ,b
2
, {J, K, L} ,b
3
,b
4
}
(c) and (b)
{c
1
,c
2
, {E,I,J} ,c
3
,c
4
,c
5
} and {b
1
,b
2
, {F} ,b
3
, {I} ,b
4
,b
5
}
(d) and (b)
{d
1
, {B,C} ,d
2
, {G} ,d
3
} and {b
1
,b
2
, {J, K, L} ,b
3
}
The detail of our proposed approach and its correspondent experimental eval-
uation is elaborated in the following sections.
3 Approaches
In this paper we propose to mine out similar patterns from pairs of process
models via a combinational approach. Before presenting this approach in detail,
we first discuss the single straightforward approach.