Enhancing the X-Processes algorithm to discover more accurate process models

• Grant number: 20/05248-4
• Support Opportunities: Regular Research Grants
• Start date: August 01, 2020
• End date: January 31, 2023
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Marcelo Fantinato
• Grantee: Marcelo Fantinato
• Host Institution: Escola de Artes, Ciências e Humanidades (EACH). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated researchers: Sarajane Marques Peres

Abstract

The combination of Business Process Management (BPM) and data mining established the research area known as process mining. Process mining aims to extract knowledge from data resulting from business process execution. The data to be mined is usually formed by event logs produced by information systems automating the organizations' business processes. Process discovery is one of the most studied and challenging types of process mining. A major challenge is to discover accurate process models, i.e., those with an appropriate balance between completeness (or recall) and precision. The X-Process algorithm was recently developed to face this challenge through genetic algorithms, using an island-based distributed approach. Although evolutionary computation is an appropriate computational approach for searching between different solutions while optimizing a specific fitness function, X-Process still needs enhancements to provide better accuracy. Therefore, the main goal of this research project is to enhance the X-Process algorithm to enable it to discover more accurate process models, by optimizing the measure F-score between completeness and precision. In addition, aiming to deliver more appropriate results, other features of this algorithm need to be improved, such as implementing additional crossover and mutation operators, using (IEEE/ISO) standard formats for input and output of the algorithm and ensuring the discovered process models are sound. Considering all the planned improvements for the X-Process algorithm, we expect to obtain the discovery of process models more meaningful for the business when compared to state-of-the-art process discovery algorithms. (AU)

Scientific publications

(4)

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

FANTINATO, MARCELO; PERES, SARAJANE MARQUES; REIJERS, HAJO A.; IEEE COMP SOC. X-Processes: Discovering More Accurate Business Process Models with a Genetic Algorithms Method. 2021 IEEE 25TH INTERNATIONAL ENTERPRISE DISTRIBUTED OBJECT COMPUTING CONFERENCE (EDOC 2021), v. N/A, p. 10-pg., 2021-01-01. (17/26491-1, 20/05248-4)

NEUBAUER, THAIS RODRIGUES; DA SILVA, VALDINEI FREIRE; FANTINATO, MARCELO; PERES, SARAJANE MARQUES; XAVIER-JUNIOR, JC; RIOS, RA. Resource Allocation Optimization in Business Processes Supported by Reinforcement Learning and Process Mining. INTELLIGENT SYSTEMS, PT I, v. 13653, p. 16-pg., 2022-01-01. (20/05248-4)

FARIA JUNIOR, ELIO RIBEIRO; NEUBAUER, THAIS RODRIGUES; FANTINATO, MARCELO; PERES, SARAJANE MARQUES; MONTALI, M; SENDEROVICH, A; WEIDLICH, M. Clustering Analysis and Frequent Pattern Mining for Process Profile Analysis: An Exploratory Study for Object-Centric Event Logs. PROCESS MINING WORKSHOPS, ICPM 2022, v. 468, p. 13-pg., 2023-01-01. (20/05248-4)

FANTINATO, MARCELO; PERES, SARAJANE MARQUES; REIJERS, HAJO A.. X-Processes: Process model discovery with the best balance among fitness, precision, simplicity, and generalization through a genetic algorithm. INFORMATION SYSTEMS, v. 119, p. 25-pg., 2023-07-10. (20/05248-4, 17/26491-1)