文档介绍：Choose initial k “seed” events from E
Determine a star for each seed against the other seed events
By appropriately modifying and plexes from stars,
construct a disjoint cover of E that optimizes the criterion LEF
a
第七章观察与发现学习

Given: E-a set of data events k-the number of clusters LEF-the clustering quality criterion
Is the clustering
quality improving?
Choose k new seeds which
Are central events
Choose k new seeds which
are “border” events
a
Y
N
Is the termination
Criterion satisfied?
END
Y
N
a
b
c
e3
a
e4
b
c
e5
a
e6
b
e7
e8
c
e9
e10
0
1
2
0
1
2
0
1
2
e1
e2
X1 X2
0
1
2
X4
X3
0 1 2
Event
X1
X2
X3
X4
e1
0
a
0
1
e2
0
b
0
0
e3
0
c
1
2
e4
1
a
0
2
e5
1
c
1
1
e6
2
a
1
0
e7
2
b
0
1
e8
2
b
1
2
e9
2
c
0
0
e10
2
c
2
2
d
f
a
b
c
K=2 ; LEF-sparseness, Complexity; Termination criterion:base=2,probe=2
Iteration 1
Step 1:
Select seed: e1,e2
Step 2:
Produce Stars: RG(e1|e2,m) RG(e2|e1,m) m=5
RG(e1|e2,m)={[x2=a][x3=0∨1],[X4=1 ∨2]}
RG(e2|e1,m)={[x2=b ∨c],[x4=0 ∨2]}
Generalize:
RG(e1|e2,m)={[x2=a][x3≦1],[X4=1∨2]}
RG(e2|e1,m)={[x2=f],[x4=0∨2]}
Step 3:
Evaluation and Modification(disjoint)
plex 1: [x2=a][x3≦1] 15 2
Complex 2: [x2=f] 47 1
62 3
(b) Complex 1: [x4=1∨2]
Complex 2: [X2=f]

(c) Complex 1: [x2=a][x3≦1]
Complex 2: [X4=0∨2]
(d) Complex 1: [x4=1∨2]
Complex 2: [x4=0∨2]
Step 4:
The termination criterion is tested
Step 5:select new seeds
{e1,e4,e6} {e2,e3,e5,e7,e8,e9,e10}
Central events: e4,e8
Iteration 2
Step 2:
Produce satrs RG(e4|e8,m ), RG(e8|e4,m)
RG(e4|e8,m)={[x2=a][x3≦1],[x1≦1][x3 ≦1],[x3=0]}
RG(e8|e4,m)={[x1=2],[x2=f],[x3≧1]}
plex 1: [x1≤1][x3≤1] 31 plex 2: [x1=2] 22 1
53 3
Step 4:
Termination criterion is tested (the last of the base iterations)
Step 5:
{e1,e2,e3,e4,e5} {e6,e7,e8,e9,e10}
New seeds: e1,e8
Iteration 3
The iteration produces the same clustering as iteration1
Step 4: Termination criterion is tested (the first of the t