Sub-process discovery: opportunities for process diagnostics

579 views

Published on

Raykenler Yzquierdo Herrera, Rogelio Silverio Castro, Manuel Lazo Cortes, Sub-process discovery: opportunities for process diagnostics

Published in: Business, Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
579
On SlideShare
0
From Embeds
0
Number of Embeds
7
Actions
Shares
0
Downloads
16
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Sub-process discovery: opportunities for process diagnostics

  1. 1. Agenda 1 Process Mining 2 Process diagnostics 3 The proposal 4 Proposal application2
  2. 2. Introduction Event logs Enterprise3
  3. 3. Introduction. Process MiningEvent logs Start Register order Prepare shipment (Re)send bill Ship goods Contact Receive payment customer Archiv e order End Mining techniques4 Ana Karla Alves de Medeiros and Van der Aalst
  4. 4. Process Mining. DiscoveryEvent logs Process Model Start Register order Prepare shipment (Re)send bill Ship goods Contact Receive payment customer Archiv e order Mining Techniques EndMost business processes in real life are not strictly ruled by theinformation systems on the background5
  5. 5. Diagnostic questions What is the most common (likely) process behavior that is executed? Where do process instances deviate and what do they have in common? Are there any common patterns of execution in the traces? What are the contexts in which an activity or a set of activities is executed in the event log? What are the process instances that share/capture a desired behavior either exactly or approximately? Are there particular patterns in the process?6 R. P. J. Chandra Bose, W. M. P. van der Aalst
  6. 6. Process diagnostics Event logs Holistic view of the process7 R. P. J. Chandra Bose, W. M. P. van der Aalst
  7. 7. Process diagnostics. Techniques. Dotted chart analysis Shortcomings Trace alignment Conformance checking Fuzzy miner Tandem arrays Fuzzy Clustering8 Yzquierdo R. (2012).
  8. 8. Approach9 Yzquierdo R. (2012).
  9. 9. 1. Trace alignment Matrix AEvent Log10 R. P. J. Chandra Bose, W. M. P. van der Aalst
  10. 10. 2. Pre-processing aligned traces  Incomplete cases  Noise  To assure each column is occupied by a single task 11
  11. 11. 3. Determining the tree of building blocks 12
  12. 12. 3. Determining the tree of building blocks 13
  13. 13. 3. Determining the tree of building blocks Sequence 14
  14. 14. 3. Determining the tree of building blocks Sequence XOR Parallelism 15
  15. 15. Tree of building blocks Sequence16
  16. 16. Tree of building blocks XOR17
  17. 17. Tree of building blocks Parallelism18
  18. 18. Tree of building blocks Loop19
  19. 19. Tree of building blocks Hidden sequence20
  20. 20. Applying the proposalNational Identification Unique System (SUIN)Module: Management of RolesCases: 31Events: 804Event classes: 52Event types: 3 21
  21. 21. Applying the proposalNational Identification Unique System (SUIN)Module: Managing Roles Are there significantCases: 31 failures in thisEvents: 804 process?Event classes: 52 What are the causes?Event types: 3 22
  22. 22. Applying the proposalNational Identification Unique System (SUIN)Module: Managing Roles Are there significantCases: 31 failures in thisEvents: 804 process?Event classes: 52 What are the causes?Event types: 3 23
  23. 23. Applying the proposal How the process endDiagnosis:• Problems: Edit Rol and Delete Rol events• Misuse of the system by an administrator
  24. 24. Discussion The technique developed allows detecting interesting patterns and provides a holistic view of the process The proposal also allows detection of sub- processes that compose the analyzed process The detected sub-processes enclose anomalies and interesting patterns It combines the cases and activities occurrence frequency analysis with the staged analysis from correctly structured sequence events on sub- processes25
  25. 25. Discussion An important contribution is that the anomalies detected can be framed in a context. The developed tool was also applied to analyze the process “Check Management” in the bar Gulf View and the restaurant Aguiar, both places belonging to the National Hotel (Cuba). Main characteristics of the process for both event logs, which supported the auditing of the process, were identified26
  26. 26. Conclusion Process diagnostics can be useful for detecting patterns and anomalies in the analyzed process This proposal allows to segment the aligned traces and to form representative groups of sub- processes that compose the analyzed process The obtained tree of building blocks reflects the hierarchical organization that is established between the sub-processes, considering main execution patterns The technique developed is a complement of the techniques developed in Process diagnostics27
  27. 27. Thanks

×