Process mining chapter_06_advanced_process_discovery_techniques
0 likes269 views
The document discusses advanced process discovery techniques. It begins by describing the challenges of process discovery, including the need for models to be able to replay event logs while avoiding overfitting or underfitting the logs. It then provides examples of algorithmic techniques like the heuristic miner and genetic process mining. Region-based process mining is also introduced. The document discusses characteristics of different process discovery algorithms and provides examples to illustrate concepts like heuristic mining, genetic operations, and region-based mining.
![Past with multiset abstraction
[a,e]
d
[a,d,e]
e [a,b]
a b
[] [a]
c c
b d
[a,c] [a,b,c] [a,b,c,d]
PAGE 31](https://image.slidesharecdn.com/processminingchapter06advancedprocessdiscoverytechniques-121219213521-phpapp01/85/Process-mining-chapter_06_advanced_process_discovery_techniques-32-320.jpg)
![Example
d
e
[a,e] [a,d,e]
[ a,b]
a b
[] [a] c
c
b d
[a,c] [a,b,c] [a,b,c,d]
b
a p1 e p3 d
start end
p2 c p4
PAGE 34](https://image.slidesharecdn.com/processminingchapter06advancedprocessdiscoverytechniques-121219213521-phpapp01/85/Process-mining-chapter_06_advanced_process_discovery_techniques-35-320.jpg)
Process mining chapter_06_advanced_process_discovery_techniques
- 1. Chapter 6 Advanced Process Discovery Techniques prof.dr.ir. Wil van der Aalst www.processmining.org
- 2. Overview Chapter 1 Introduction Part I: Preliminaries Chapter 2 Chapter 3 Process Modeling and Data Mining Analysis Part II: From Event Logs to Process Models Chapter 4 Chapter 5 Chapter 6 Getting the Data Process Discovery: An Advanced Process Introduction Discovery Techniques Part III: Beyond Process Discovery Chapter 7 Chapter 8 Chapter 9 Conformance Mining Additional Operational Support Checking Perspectives Part IV: Putting Process Mining to Work Chapter 10 Chapter 11 Chapter 12 Tool Support Analyzing “Lasagna Analyzing “Spaghetti Processes” Processes” Part V: Reflection Chapter 13 Chapter 14 Cartography and Epilogue Navigation PAGE 1
- 3. Process discovery supports/ “world” business controls processes software people machines system components organizations records events, e.g., messages, specifies transactions, models configures etc. analyzes implements analyzes discovery (process) event conformance model logs enhancement PAGE 2
- 4. Challenge “able to replay event log” “Occam’s razor” fitness simplicity process discovery generalization precision “not overfitting the log” “not underfitting the log” PAGE 3
- 5. Observing a stable process infinitely long frequent all behavior behavior trace in (including noise) event log PAGE 4
- 6. Target model target model PAGE 5
- 7. Non-fitting model non-fitting model PAGE 6
- 8. Overfitting model overfitting model PAGE 7
- 9. Underfitting model underfitting model PAGE 8
- 10. Characteristics of process discovery algorithms • Representational bias − Inability to represent concurrency − Inability to deal with (arbitrary) loops − Inability to represent silent actions − Inability to represent duplicate actions − Inability to model OR-splits/joins − Inability to represent non-free-choice behavior − Inability to represent hierarchy • Ability to deal with noise • Completeness notion assumed • Approach used (direct algorithmic approaches, two- phase approaches, computational intelligence approaches, partial approaches, etc.) PAGE 9
- 11. Examples • Algorithmic techniques • Alpha miner • Alpha+, Alpha++, Alpha# • FSM miner • Fuzzy miner • Heuristic miner • Multi phase miner • Genetic process mining • Single/duplicate tasks • Distributed GM • Region-based process mining • State-based regions • Language based regions • Classical approaches not dealing with concurrency • Inductive inference (Mark Gold, Dana Angluin et al.) • Sequence mining PAGE 10
- 12. Heuristic mining • To deal with noise and incompleteness. • To have a better representational bias than the α algorithm (AND/XOR/OR/skip). • Uses C-nets. b check policy a c e register check close claim damage case d consult expert PAGE 11
- 13. Example log; problem α algorithm p5 b a p1 d p3 e start end p2 c p4 PAGE 12
- 14. Taking into account frequencies PAGE 13
- 15. Dependency measure PAGE 14
- 16. Example PAGE 15
- 17. Lower threshold (2 direct successions and a dependency of at least 0.7) 5(0.83) b 11(0.92) 11(0.92) a c e 11(0.92) 11(0.92) 13(0.93) 13(0.93) d 4(0.80) PAGE 16
- 18. Higher threshold (5 direct successions and a dependency of at least 0.9) b 11(0.92) 11(0.92) a c e 11(0.92) 11(0.92) 13(0.93) 13(0.93) d PAGE 17
- 19. Learning splits and joins 5 20 b 20 21 5 20 20 5 20 20 20 20 a c e 40 20 21 20 40 13 13 13 13 13 13 d 4 17 4 4 PAGE 18
- 20. Alternative visualization 5 20 b 20 21 5 20 20 5 20 20 20 20 a c e 40 20 21 20 40 13 13 13 13 13 13 d b 4 17 4 4 AND AND a c e d PAGE 19
- 21. Characteristics of heuristic mining • Can deal with noise and therefore quite robust. • Improved representational bias. • Split and join rules are only considered locally (therefore most of the discovered model are not sound and require repair actions). PAGE 20
- 22. Genetic process mining create initial population event log mutation next generation compute fitness elitism termination tournament children crossover select best parents individual “dead” individuals PAGE 21
- 23. Design decisions • Representation of individuals • Initialization • Fitness function • Selection strategy (tournament and elitism) • Crossover create initial population • Mutation event log mutation next generation compute fitness elitism termination tournament children crossover select best parents individual “dead” individuals PAGE 22
- 24. Example: crossover b b examine examine thoroughly thoroughly g g pay pay c c compensation compensation a e a e examine examine start register casually decide end start register casually decide end request request h h d d reject reject check ticket request check ticket request f f reinitiate reinitiate request request b b examine examine thoroughly thoroughly g g pay pay c c compensation compensation a e a e examine examine start register casually decide end start register casually decide end request request h h d d reject reject check ticket request check ticket request f f reinitiate reinitiate request request PAGE 23
- 25. Example: mutation remove place b b examine examine thoroughly thoroughly g g pay pay c c compensation compensation a e a e examine examine start register casually decide end start register casually decide end request request h h d d reject reject check ticket request check ticket request f f reinitiate reinitiate request added arc request PAGE 24
- 26. Characteristics of genetic process mining • Requires a lot of computing power. • Can be distributed easily. • Can deal with noise, infrequent behavior, duplicate tasks, invisible tasks, etc. • Allows for incremental improvement and combinations with other approaches (heuristics post-optimization, etc.). PAGE 25
- 27. Region-based mining • Two types of regions theory: − State-based regions − Language-based regions • All about discovering places (like in the α algorithm)! a1 b1 a2 b2 ... p(A,B) ... am bn A={a1,a2, … am} B={b1,b2, … bn} PAGE 26
- 28. State-based regions Two steps: 1.Discover a transition system (different abstractions are possible) 2.Convert transition system into an “equivalent” Petri net. PAGE 27
- 29. Step 1: learning a transition system current state trace: abcdcdcde faghhhi past future past and future • past, future, past+future • sequence, multiset, set abstraction • limited horizon to abstract further • filtering e.g. based on transaction type, names, etc. • labels based on activity name or other features PAGE 28
- 30. Past without abstraction (full sequence) c d ‹a,b› ‹a,b,c› ‹a,b,c,d› b a e d ‹› ‹a› ‹a,e› ‹a,e,d› c b d ‹a,c› ‹a,c,b› ‹a,c,b,d› PAGE 29
- 31. Future without abstraction a b ‹c,d› ‹a,b,c,d› ‹b,c,d› c a e d ‹a,e,d› ‹e,d› ‹d › ‹› b a c ‹b,d› ‹a,c,b,d› ‹c,b,d› PAGE 30
- 32. Past with multiset abstraction [a,e] d [a,d,e] e [a,b] a b [] [a] c c b d [a,c] [a,b,c] [a,b,c,d] PAGE 31
- 33. Only last event matters for state ‹e› e d a b ‹ b› d ‹› ‹a › c b ‹d› c d ‹c› PAGE 32
- 34. Step 2: constructing a Petri net using regions a = enter b d b = enter a e c = exit d = exit f d e = do not cross e f = do not cross e f c a R a c e f pR b d PAGE 33
- 35. Example d e [a,e] [a,d,e] [ a,b] a b [] [a] c c b d [a,c] [a,b,c] [a,b,c,d] b a p1 e p3 d start end p2 c p4 PAGE 34
- 36. Language based regions f c1 a1 b1 e c d pR a2 b2 X Y Region R = (X,Y,c) corresponding to place pR: X = {a1,a2,c1} = transitions producing a token for pR, Y = {b1,b2,c1} = transitions consuming a token from pR, and c is the initial marking of pR. PAGE 35
- 37. Based idea: enough tokens should be present when consuming A place is feasible if it can be added without f c1 disabling any of the traces in the event log. a1 b1 e c d pR a2 b2 X Y PAGE 36
- 38. Example PAGE 37
- 39. Regions PAGE 38
- 40. Model a p5 d c p1 p2 p3 p4 b e p6 PAGE 39
- 41. Characteristics of region-based mining • Can be used to discover more complex control-flow structures. • Classical approaches need to be adapted (overfitting!). • Representational bias can be parameterized (e.g., free-choice nets, label splitting, etc.). • Problems dealing with noise. PAGE 40
- 42. Other approaches, e.g. fuzzy mining PAGE 41
- 43. Evaluating the discovered process Fitness: Is the event log possible according to the model? Precision: Is the model Generalization: Is the model not underfitting (allow for not overfitting (only allow for too much)? the “accidental” examples)? Structure: Is this the simplest model (Occam's Razor)? PAGE 42