Process mining

Process mining or Process Mining (PM) is a technique process management to analyze business process according to an event log.

Summary

[ hide ]

1 Introduction
2 Definition
3 Emergence and development
4 Objectives
5 Description
6 Main Tools and techniques
- 1 Diffuse Mining
- 2 Avg
- 3 Disc
7 Basic Types of Process Mining
- 1 Discovery
- 2 Verification of Conformity
- 3 Improvement
8 Feasibility of Process Mining
9 Benefits of applying Process Mining
10 Advantages
11 Disadvantages
12 Applications
13 Data Mining versus Process Mining
14 Importance
15 Sources

Introduction

As a way to make their operations more efficient and remain competitive on a world increasingly interconnected and globalized, organizations are paying increasing attention to business processes running in your organization. The number of information systems that support business processes is increasing, from which arises the natural idea of taking advantage of the information that is registered in these systems to learn about the historical behavior of the processes and thus detect sources of problems and opportunities. From this last idea the discipline arises Process Mining, a discipline that seeks to analyze the information that information systems record about the business processes that they support, in order to understand, monitor, analyze and improve said processes.

Definition

Process Mining is a process management technique that allows you to analyze business processes according to an event log. Through this activity you want to extract knowledge from the event logs of the processes stored by the systems. This knowledge implies achieving the trace of the processes under study , including information on the actors who carry it out, the times involved, among other things. It is a research discipline that is located between computational intelligence and Data Mining on the one hand, and process modeling and analysis on the other, makes it possible to understand how processes are actually executed in thesystem .

Emergence and development

This concept was defined in the Process Mining Manifesto, launched by the worldwide group of experts investigating the MP, known as the IEEE Task Force on Process Mining, in the 2012 and supported by 53 organizations, with the contribution of 77 experts on the subject. It is aimed at promoting the topic of Process Mining. Furthermore, by defining a set of guiding principles and listing important challenges, the manifesto serves as a guide for software developers , scientists, consultants, business managers, and end users. The objective is to increase the maturity of Process Mining as a new toolto improve the (re) design, control, and support of operational business processes. There are two main reasons for the growing interest in Process Mining: on the one hand, more and more events are recorded, providing detailed information about the history of the processes and on the other hand, there is a need to improve and support business processes in competitive and rapidly changing environments.

objectives

Take control of the processes
Allow the discovery of processes, controls, information and organizational structures based on the event logs.
Its application helps identify necks bottle , anticipate problems, record policy violations, recommend countermeasures, and streamline processes.

Description

Process Mining is generally used when there is no formal description of the processes, or when the existing information is of poor quality . For example, a patient’s care record in a hospitalIt can be analyzed through Process Mining to discover models that describe patient flow. On the other hand, event logs can also be used to compare them with probabilistic models to contrast reality with the proposed model. Being the discovery of process models an initial stage, the next step of Process Mining is to achieve an improvement in the studied processes by detecting bottlenecks in the process, better routes to fulfill the same purpose, early detection of congestion, among other things.

Main tools and techniques

Process Mining is made up of a series of tools and techniques based on Data Mining to analyze business processes whose record of actual execution events are available in information systems.

Diffuse Mining

Among the Process Mining techniques is the Diffuse Mining technique that provides reliable and trustworthy results for a complex data set, facilitating analysis and decision-making on them. It is the first algorithmthat deals directly with problems when there is a large accumulation of activities and behavior in unstructured processes. It makes use of a mixture of techniques for extraction and grouping in order to represent a process that is understandable by process analysts. It is a technique that allows obtaining models with an adequate pre-adjustment and over-adjustment balance with respect to event or trace records, it is obtained practically immediately, which speeds up the obtaining of results and gives us great versatility to when carrying out analysis and decision-making, in order to optimize and improve the processes to which this technique is applied.

Prom

It is an open source academic tool for Process Mining. It requires experience in Process Mining since it is not focused on usability and is not endorsed by a commercial organization therefore it has the common advantages and disadvantages for open source software. Enables discovery process, compliance checking, social media analysis, organizational mining, decision mining.

Disk

Developed by Fluxicon in 2009 , it is a great academic process mining tool that helps organizations support the control of their processes and simplify work.It is ideal for dealing with large event logs and makes complex model conversion and filtering easy. It is based on the Diffuse Mining technique although other techniques have been developed. The result that is obtained is reliable and trustworthy for a data set of arbitrary complexity, in addition that it can be efficiently operated and understood by domain experts without prior experience in Process Mining. Disco is also fully compatible with the ProM 5 and 6 academic toolkits. By importing and exporting the standard MXML and XES event log formats, advanced users can seamlessly move back and forth between Disco and ProM if they want to benefit of the new research technologies developed in the academic field.

Basic Types of Process Mining

Discovery

An event log is used to produce a model without using a-priori. Process discovery is the most outstanding process mining technique. It is surprising to many organizations to see that existing techniques are truly capable of discovering actual processes based purely on event log execution samples.

Conformity Verification

They need an event log and model as input. The output consists of diagnostic information showing the differences and common elements between the model and the event log. Here an existing process is compared with an event log of the same process, to verify if the reality, according to the log, is equivalent to the model and vice versa.

Improvement

The aim is to extend or improve an existing process model with the information of the real process stored in an event log. They also need an event log and template as input. The output is an improved or extended model

Feasibility of Process Mining

There is in the companies great detailed information on the execution of their high quality processes:

– Auditors do not need to rely on a small set of process data samples.

The support of Process Mining allows evaluating all the events of a business process:

– It is done in its own execution

Therefore, traceability is important:

– The story cannot be modified, remade or hidden.

Benefits of applying Process Mining

Discover a model in fraction of time versus the traditional method:

– Minimum design time

Real design, without political influences or forgetfulness of functions:

– Efficient conformity check

Filter irrelevant situations and customize the required analysis:

– Minimum analysis time

Simulation of possible redesigns in progress:

– The parameters are real

Control and strategic redesign of processes:

– Focusing on what’s important

Advantage

This relatively young discipline when carrying out its analyzes on the evidence of the execution of the processes offers numerous advantages with respect to the traditional techniques of analysis and process improvement, since it offers results in less time and with greater reliability.
Unlike traditional process modeling, which uses tools and where the human factor and the conception of what to do intervenes, Process Mining shows what really happens. This allows a more analytical study of the processes and operation of a company , based on events that have already occurred and that show the reality of the execution of the processes that are stored as event logs.

Disadvantages

MP techniques are used to discover the business process model from the logs. These techniques do not properly handle some situations like the ones shown below:

Noise: The recorded data may be incorrect or incomplete, creating problems when the data is mined. In other words, there is infrequent or exceptional behavior.
Hidden tasks: Tasks that exist but cannot be found in the data.
Duplicate tasks: Two process nodes can refer to the same process model.
Builders without free choice: Controlled options that depend on selections made elsewhere in the process model.
Mined cycles: A process can be run multiple times, the cycles may simply be involving one or more events, or they may be more complex.
Different perspectives: Process events can be added with additional information for specific purposes.

Applications

Process Mining follows the known possibilities of business process engineering and also contributes to business process modeling:

Process analysis: Filters, sorts and compresses the log files to facilitate connections between the operation of the processes.
Process design: can be maintained thanks to the information stored in the event logs. This allows us to determine which is the process model used empirically.
Process validation: uses the results of Process Mining, based on event logs. to trigger some operation ..

Data Mining versus Process Mining

Process mining is Data Mining but with a strong vision of business processes.
Some of the more traditional Data Mining techniques can be used in the context of Process Mining.
Some of the new techniques are developed to execute Process Mining (extraction and / or discovery of process models)
The goal of Process Mining is to discover, monitor and improve real processes through the extraction of knowledge from the information logs.

Importance

Process Mining is of vital importance for the Information systems because it allows an objective analysis of the processes based on the execution of the activities of the organization.