Genetic Process Mining and the Prom Framework

1
Genetic Process Mining(and the Prom Framework!)

• Ana Karla Alves de Medeiros
• Eindhoven University of Technology
• Department of Information Systems
• a.k.medeiros_at_tm.tue.nl

2
Outline

• Process Mining
• Motivation
• Current Techniques
• Contribution
• Genetic Algorithms
• Experiments and Results
• Conclusion and Future Work

3
Outline

• Process Mining
• Motivation
• Current Techniques
• Contribution
• Genetic Algorithms
• Experiments and Results
• Conclusion and Future Work

4
Motivation Typical way to deploy a system
5
Motivation Typical way to deploy a system

• Time consuming
• Paper procedures
• Meetings
• Error prone
• Different people have different views about a
  same process
• Information about the process may be incomplete

6
Motivation More possible cases in the log
7
Motivation Process Mining
Mined Model
Log
Mining Techniques
8
Motivation Process Mining

• Before deployment
• Objective picture of how the process has been
  executed
• After deployment
• Feedback mechanism

9
Current Techniques vs Constructs

• Sequence
• Splits
• Joins
• Loops
• Non-free Choice
• Invisible Tasks
• Duplicate Tasks

10
Current Techniques vs Constructs

• Sequence
• Splits
• Joins
• Loops
• Non-free Choice
• Invisible Tasks
• Duplicate Tasks

11
Current Techniques vs Constructs

• Sequence
• Splits
• Joins
• Loops
• Non-free Choice
• Invisible Tasks
• Duplicate Tasks

12
Current Techniques vs Constructs

• Sequence
• Splits
• Joins
• Loops
• Non-free Choice
• Invisible Tasks
• Duplicate Tasks

13
Current Techniques vs Constructs The ?-algorithm
14
Current Techniques

• Problematic structural constructs
• Non-free choice, invisible tasks, duplicate tasks
• Noise
• Wrongly logged traces
• Exceptional situations

• Can we develop a mining algorithm that is able to
  tackle all structural constructs and noisy logs
  at the same time?

15
Outline

• Process Mining
• Motivation
• Current Techniques
• Contribution
• Genetic Algorithms
• Experiments and Results
• Conclusion and Future Work

16
Genetic Algorithms

• Global approach

17
Genetic Process Mining (GPM)
Aim Use genetic algorithm to tackle non-free
choice, invisible tasks, duplicate tasks and
noise.
Internal Representation
Fitness Measure
Genetic Operators
18
Genetic Process Mining (GPM)

• Demo
• Genetic Miner plug-in
• Tool at www.processmining.org

19
GPM Fitness Measure
20
GPM Fitness Measure
21
GPM Fitness Measure
Overgeneral solution
Punish for the amount of enabled tasks during the
parsing!
22
GPM Fitness Measure
Start
Overspecific solution
Get Ready
Travel by Car
Travel by Train
Beta Event Starts
Give a Talk
Visit Brewery
Have Dinner
Go Home
Punish for the amount of duplicate tasks with
common input/output tasks!
Travel by Train
Pay Parking
Travel by Car
End
23
Outline

• Process Mining
• Motivation
• Current Techniques
• Contribution
• Genetic Algorithms
• Experiments
• Conclusion and Future Work

24
Experiments

• Simulation to generate the logs

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
Original Process
Logs
Genetic Miner
Mined Process
Can the mined process generate all the behavior
in the log? How close is the behavior of the
mined process to the original one?
25
Outline

• Process Mining
• Motivation
• Current Techniques
• Contribution
• Genetic Algorithms
• Experiments and Results
• Conclusion and Future Work

26
Conclusion and Future work

• Genetic algorithms can be used to mine process
  models
• Global approach
• Robust to noise
• Run more experiments
• Case study
• Town hall Heusden
• Write thesis ?

27
http//www.processmining.org