The document discusses an iterative clustering technique for conceptual clustering. It begins by asking what clustering and concepts are, then describes an iterative approach to cluster extraction and navigation of a graph. The approach involves multiple instances of clustering, first for packages and then for subprograms. Issues addressed include cyclic dependencies and multiple allocation. Metrics for precision and recall are provided for the two instances.

1. Clustering technique
for conceptual cluster
Brice Govin

2. Clustering technique for
conceptual cluster
• Clustering ?
• Concept?
• An iteration to rule them all
• And in an approach bind them
• And come the hobbits and their issues

3. Clustering?

4. Clustering?

5. Clustering?

6. Clustering?

7. Concept
?

8. Concept?
Conceptual cluster

9. Clustering concepts

10. An iteration to rule them all
Extracted graph
A
E
I
B
D
F
H
G
C
D Packages
Imports

11. An iteration to rule them all
Kernel Selection
A
E
I
B
D
F
H
G
C

12. An iteration to rule them all
Navigating the graph
A
E
I
B
D
F
H
G
C

13. An iteration to rule them all
Navigating the graph
A
E
I
B
D
F
H
G
C

14. An iteration to rule them all
Navigating the graph
A
E
I
B
D
F
H
G
C

15. An iteration to rule them all
Navigating the graph
A
E
I
B
D
F
H
G
C
G

16. An iteration to rule them all
Difference between traditional methods
and ours
A
E
I
B
D
F
H
G
C
dist(A,F)
similarity(A,H)

17. And in an approach
bind them

18. And in an approach bind them
Why iterative?
Because hierarchy of clusters
Cluster
1
Cluster
1.3
Cluster
2.1
Cluster
3.1
Cluster
3.2
Cluster
2
Cluster
3
Cluster
1.2
Cluster
1.1
Cluster
2.2

19. And in an approach bind them
How it works?
First instance, packages clustering
Packages
Subprograms

20. And in an approach bind them
How it works?
First instance, packages clustering
Packages
Cluster
1
Cluster
2
Cluster
3

21. And in an approach bind them
How it works?
2nd instance, subprograms clustering
Packages
Subprograms

22. And in an approach bind them
How it works?
2nd instance, subprograms clustering
Subprograms
Cluster
1.3
Cluster
2.1
Cluster
3.1
Cluster
3.2
Cluster
1.2
Cluster
1.1
Cluster
2.2

23. And come the hobbits and their issues
Cyclic dependencies
A
E
I
B
D
F
H
G
C

24. And come the hobbits and their issues
Cyclic dependencies
Keep track of navigated nodes
A
E
I
B
D
F
H
G
C

25. And come the hobbits and their issues
Multiple allocation
A
E
I
B
D
F
H
G
C

26. Talk about numbers
1st Instance: packages clustering
2nd Instance: subprograms clustering
Precision: 98%
Recall: 88%
Precision: ~60%
Recall: ~50%%

27. What next?