Decision Support Using Belief Network Constructed from Business Process Event Log

Article type: Research Article

Authors: Savickas, Titas^{1; *} | Vasilecas, Olegas^{1; 2}

Affiliations: [1] Information Systems Department, Vilnius Gediminas Technical University, Lithuania. E-mail: titas.savickas@vgtu.lt | [2] Institute of Mathematics and Informatics, Vilnius University, Lithuania. E-mails: olegas.vasilecas@mii.vu.lt, olegas.vasilecas@vgtu.lt

Correspondence: [*] Corresponding author.

Abstract: Information systems contain a lot of data regarding business process execution history. Use of this data, in the form of an event log, can greatly support business process management. The paper presents an approach to construct Bayesian belief network from an event log that could facilitate decision support in business process execution. The approach is evaluated against multiple event logs by inferring data probabilities occurring in the business processes. The results show that the approach is suitable for the task and could be used in decision support with future research focused on prediction and simulation of business processes.

Keywords: event log, Bayesian belief network, decision support, probability inference

Journal: Informatica, vol. 28, no. 4, pp. 687-701, 2017