11. Workflow Mining Discovering Process Models From Event Logs
Workflow Mining
-
Upload
seyed-ziae-mousavi-mojab -
Category
Documents
-
view
220 -
download
0
Transcript of Workflow Mining
-
7/28/2019 Workflow Mining
1/24
Workflow mining:
A survey of issues and approaches
W.M.P. van der Aalst, B.F. van Dongen, J. Herbst,
L. Maruster, G. Schimm, A.J.M.M. Weijters
Eindhoven University of Technology
Presenter: Seyed Ziae Mousavi MojabWayne State University - 2013
-
7/28/2019 Workflow Mining
2/24
Agenda
Introduction Related work
Workflow mining
Workflow log
Comparison and open problems
Conclusion
-
7/28/2019 Workflow Mining
3/24
Introduction
Data mining:The computational process of discovering patterns in large data
sets involving methods at the intersection of artificial
intelligence, machine learning, statistics, and database systems
The process of
analyzing data from
different
perspectives andsummarizing it into
useful information
-
7/28/2019 Workflow Mining
4/24
Introduction
Workflow Life-Cycle: to construct a workflowmodel
to deal with limitation andparticularities of the
workflow management
system
to run and execute aworkflow and collect
diagnostic information
to provide input for thedesign phase
-
7/28/2019 Workflow Mining
5/24
Introduction
The Goal ofworkflow mining:
To reverse the process and collect data at runtime to supportworkflow design and analysis.
-
7/28/2019 Workflow Mining
6/24
Introduction
Workflow mining techniques:
o To create a feedback loop to adapt the workflow modelto change circumstances and detect imperfections of the
design
Reengineering Vs. Improvements
-
7/28/2019 Workflow Mining
7/24
Related work
Process mining approaches (investigated by Cook & Wolf):
o Neural Networks Approacho Pure Algorithmic Approach
o Markovian Approach
-
7/28/2019 Workflow Mining
8/24
Related work
Process mining approaches:
o Neural Networks are non-linear predictive models that learn through
training and resemble biological neural networks in
structure
-
7/28/2019 Workflow Mining
9/24
Related work
Process mining approaches:
o Pure Algorithmic Approach builds a finite state machine (FSM), where states are
fused if their futures (in terms of possible behavior in
the next k steps) are identical.
-
7/28/2019 Workflow Mining
10/24
Related work
Process mining approaches:
o Markovian Approach uses a mixture of algorithmic and statistical methods
and is able to deal with noise
-
7/28/2019 Workflow Mining
11/24
Related work
Final goal of discovery approaches:
o we want to be able to generate a concrete Petri net ratherthan a set of dependency relations between events.
Petri net: is one of several mathematical modeling
languages for the description of distributed systems
-
7/28/2019 Workflow Mining
12/24
Workflow Mining
Workflow mining:
o The goal of workflow mining is to extract informationabout processes from transaction logs.
o Instead of starting with a workflow design, we start by
gathering information about the workflow processes asthey take place.
-
7/28/2019 Workflow Mining
13/24
Workflow Mining
Workflow mining:
o We assume: each event refers to a task (i.e: step in a workflow)
each event refers to a case (i.e: workflow instance)
events are totally ordered
-
7/28/2019 Workflow Mining
14/24
Workflow Mining
Workflow mining:
o Definition: The term workflow (process) mining refers to
methods for distilling a structured process
description from a set of real execution.
-
7/28/2019 Workflow Mining
15/24
Workflow Mining
Workflow mining:
o The challenge of workflow mining is to derive goodworkflow models with as little information as possible
-
7/28/2019 Workflow Mining
16/24
Workflow Mining
-
7/28/2019 Workflow Mining
17/24
Workflow Mining
-
7/28/2019 Workflow Mining
18/24
Workflow Logs
Logs:
o The XML format is used as input for the analysis tools
o The DTD specifies the syntax of a workflow log
o It's a well-formed and valid XML file with top element
WorkFlow_log
o consists of (optional) information about the source
program and information about one or more workflow
processes.
o Both processes and cases have an id and a description ...
-
7/28/2019 Workflow Mining
19/24
Workflow Logs
-
7/28/2019 Workflow Mining
20/24
Comparison & Open Problems
Process mining tools:
o EMiT: can discover the underlying process model and
represent it in terms of a Petri net
o Little Thumb: is a tool that attempts to induce a
workflow model from a possibly noisy and incomplete
workflow log
o InWoLvE: to deal with duplicate tasks
o Process Miner: exploiting the properties of block-
structured workflows through rewriting rules
o ...
-
7/28/2019 Workflow Mining
21/24
Comparison & Open Problems
-
7/28/2019 Workflow Mining
22/24
Comparison & Open Problems
Open problems:
o Few of the tools exploit timing information
o Only can deal with basic routing constructs such as basic
parallelism and basic loops
o Problem of noise
o ...
-
7/28/2019 Workflow Mining
23/24
Conclusion
Summary:
o presented an overview of the various problems,
techniques, tools, and approaches for workflow mining
o compared the process mining tools by focusing on nine
aspects (Structure, Time, Basic parallelism, ...).
o revealed differences and also pointed out problems that
need to be tackled.
-
7/28/2019 Workflow Mining
24/24
Thank You!