Loading...
Publication
Descoberta automática de práticas de trabalho com técnicas de mineração de processos
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.73 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A passagem para um mundo cada vez mais digital tem vindo a concretizar-se com
o avanço da tecnologia a ocorrer de forma exponencial. Com ela tem crescido a
necessidade de se analisar a imensidão de dados que empresas e organizações produzem
a fim de identificar padrões, melhorias operacionais e tendências.
A literatura tem-se dedicado ao estudo de processos de negócio e o Process
Mining tem vindo a ganhar notoriedade como uma disciplina imprescindível ao
crescimento de uma organização no que respeita à forma como analisa os seus dados, que
passam a ser o Storyteller das mesmas. A mineração de processos dá respostas a questões
que não ficam resolvidas pelas formas tradicionais de análise de dados, permitindo extrair
conhecimento através de registos de eventos e assim descobrir novos modelos de
processos, controlar e melhorar os processos já existentes.
O estudo de caso sobre o qual este trabalho é desenvolvido debruça-se sobre o
contexto organizacional real, na identificação de práticas de trabalho a partir de registo
de ações relacionadas ao desenvolvimento de aplicações de web para um banco comercial.
Como objetivo de estudo, procurou-se realizar uma análise comparativa de vários
algoritmos de Process Mining disponíveis na ferramenta ProM.
A análise realizada teve por base encontrar o algoritmo mais apropriado às
especificações dos dados do caso de estudo, através de medidas de qualidade presentes
na verificação de conformidade.
Numa primeira fase, os dados foram extraídos, mapeados e importados para a
ferramenta ProM e seguidamente aplicaram-se algoritmos de descoberta de processo. O
resultado final gerou modelos de processo em rede Petri Net que foram analisados através
da verificação de conformidade. Por fim, foi possível verificar através dessas análises que
dos algoritmos testados, o que obteve melhor resultado foi o Heuristic Miner.
A passage to an increasingly digital world has been taking place with the advancement of technology taking place exponentially. With it, the need to analyze the immensity of data that essential companies and associations to identify patterns, operational improvements and trends has grown. Literature has been dedicated to the study of business processes and Process Mining has been gaining notoriety as an essential discipline for the growth of an organization with regard to the way it analyzes its data, which becomes the Storyteller of the previous ones. Process mining provides answers to questions that are not resolved by traditional forms of data analysis, allowing you to extract knowledge through event records and thus discover new process models, and control existing processes. The case study on which this work is developed focuses on the real organizational context, in the identification of work practices from the registration of actions related to the development of web applications for a commercial bank. As a study objective, it is expected to perform a comparative analysis of several Process Mining algorithms available in the ProM tool. The analysis performed was based on finding the most appropriate algorithm to the specifications of the case study data, through quality measures present in the compliance check. In a first phase, data were extracted, mapped and imported into a ProM tool, then process discovery algorithms were applied. The end result generated process models in Petri Net network that went through the compliance check. Finally, it was possible to verify through these analyzes that of the tested algorithms, the one that obtained the best result was the Heuristic Miner.
A passage to an increasingly digital world has been taking place with the advancement of technology taking place exponentially. With it, the need to analyze the immensity of data that essential companies and associations to identify patterns, operational improvements and trends has grown. Literature has been dedicated to the study of business processes and Process Mining has been gaining notoriety as an essential discipline for the growth of an organization with regard to the way it analyzes its data, which becomes the Storyteller of the previous ones. Process mining provides answers to questions that are not resolved by traditional forms of data analysis, allowing you to extract knowledge through event records and thus discover new process models, and control existing processes. The case study on which this work is developed focuses on the real organizational context, in the identification of work practices from the registration of actions related to the development of web applications for a commercial bank. As a study objective, it is expected to perform a comparative analysis of several Process Mining algorithms available in the ProM tool. The analysis performed was based on finding the most appropriate algorithm to the specifications of the case study data, through quality measures present in the compliance check. In a first phase, data were extracted, mapped and imported into a ProM tool, then process discovery algorithms were applied. The end result generated process models in Petri Net network that went through the compliance check. Finally, it was possible to verify through these analyzes that of the tested algorithms, the one that obtained the best result was the Heuristic Miner.
Description
Keywords
Process mining Heuristic iner ProM Petri Net Event log Verificação de conformidade