ISISTAN   23985
INSTITUTO SUPERIOR DE INGENIERIA DEL SOFTWARE
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Detection of sequences with anomalous behavior in a workflow process
Autor/es:
ARMENTANO, MARCELO G.; AMANDI, ANALIA A.
Lugar:
Valencia
Reunión:
Conferencia; 26th International Conference on Database and Expert Systems Applications (DEXA 2015); 2015
Institución organizadora:
Instituto Tecnológico de Informática, Universidad Politécnica de Valencia
Resumen:
A workflow process consists of an organized and repeatable pattern of activities that are necessary to complete a task, within the dynamics of an organization. The automatic recognition of deviations from the expected behavior within the workflow of an organization is crucial to provide assistance to new employees to accomplish his/her tasks. In this article, we propose a two-fold approach to this problem. First, taking the process logs as an input, we automatically build a statistical model that captures regularities in the activities carried out by the employees. Second, this model is used to track the activities performed by the employees to detect deviations from the expected behavior, according to the normal workflow of the organization. An experimental evaluation with five processes logs, with different levels of noise, was conducted to determine the validity of our approach.