Búsqueda avanzada

A Query Language for Exploring Directly-Follows Graph Collections


Visualization tools are very useful for data exploration and offer a very intuitive way to look at interesting results in data analysis. In previous work, systems have been designed that, starting from a dataset, generate and work on sets of data visualizations and find those that show desired trends automatically, thus avoiding manual exploration of many visualizations by the user. One of the most used mechanisms to obtain interesting visualizations is the use of a query-based language. However, the use of these systems in process mining is not contemplated, which would be very useful to specifically find interesting results among multiple Directly-Follows Graphs (DFGs) extracted from event logs without carrying out the typical manipulation tasks and exploring multiple DFGs. We are interested in extending an existing query-based language, adapting it to process mining. The goal of is to automatically generate DFG collections and search for visualizations that contain patterns of interest according to some queries made by the users. Thus, interesting visualizations can be found by obtaining and comparing properties of sets of DFGs generated by the system. As an advantage, this approach allows users to obtain interesting results without the need to carry out a manual exploration of a large number of visualizations using the existing process mining tools. We have carried out the evaluation of our approach by solving a challenge provided in a BPI Challenge.

Palabras Clave:

Data Exploration - directly-follows graphs - Process Mining - Queries - visualizations





Este artículo tiene una licencia de uso CreativeCommons - Reconocimiento (by)

Descarga el artículo haciendo click aquí.

Ver la referencia en formato Bibtex