Workflow Mining 1b3870

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 Mojab Wayne State University - 2013

Agenda

• • • • • •

Introduction Related work Workflow mining Workflow log Comparison and open problems Conclusion

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 and summarizing it into useful information

Introduction

• •

Workflow Life-Cycle:

• •

to construct a workflow model to deal with limitation and particularities of the workflow management system to run and execute a workflow and collect diagnostic information to provide input for the design phase

Introduction

The Goal of workflow mining: To reverse the process and collect data at runtime to workflow design and analysis.

Introduction

Workflow mining techniques:

o

To create a loop to adapt the workflow model to change circumstances and detect imperfections of the design



Reengineering Vs. Improvements

Related work

Process mining approaches (investigated by Cook & Wolf):

o Neural Networks Approach o Pure Algorithmic Approach o Markovian Approach

Related work

Process mining approaches:

o

Neural Networks  are non-linear predictive models that learn through training and resemble biological neural networks in structure

Related work

Process mining approaches:

o

Pure Algorithmic Approach  builds a finite state machine (FSM), where states are fused if their futures (in of possible behavior in the next k steps) are identical.

Related work

Process mining approaches:

o

Markovian Approach  uses a mixture of algorithmic and statistical methods and is able to deal with noise

Related work

Final goal of discovery approaches:

o

we want to be able to generate a concrete Petri net rather than a set of dependency relations between events.  Petri net: is one of several mathematical modeling languages for the description of distributed systems

Workflow Mining

Workflow mining:

o

The goal of workflow mining is to extract information about processes from transaction logs.

o

Instead of starting with a workflow design, we start by gathering information about the workflow processes as they take place.

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

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.

Workflow Mining

Workflow mining:

o

The challenge of workflow mining is to derive ‘‘good’’ workflow models with as little information as possible

Workflow Mining

Workflow Mining

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 o

o

WorkFlow_log consists of (optional) information about the source program and information about one or more workflow processes. Both processes and cases have an id and a description ...

Workflow Logs

Comparison & Open Problems

Process mining tools:

o EMiT: can discover the underlying process model and o o o o

represent it in of a Petri net Little Thumb: is a tool that attempts to induce a workflow model from a possibly noisy and incomplete workflow WoLvE: to deal with duplicate tasks Process Miner: exploiting the properties of blockstructured workflows through rewriting rules ...

Comparison & Open Problems

Comparison & Open Problems

Open problems:

o Few of the tools exploit timing information o Only can deal with basic routing constructs such as basic o o

parallelism and basic loops Problem of noise ...

Conclusion

Summary:

o presented an overview of the various problems, o o

techniques, tools, and approaches for workflow mining compared the process mining tools by focusing on nine aspects (Structure, Time, Basic parallelism, ...). revealed differences and also pointed out problems that need to be tackled.

Thank You!