Authors:
Souhail Boushaba
;
Mohammed Issam Kabbaj
and
Zohra Bakkoury
Affiliation:
Ecole Mohammadia d’Ingénieurs and Mohammed Vth University - Agdal, Morocco
Keyword(s):
Process Mining, Business Process Management, Process Discovery, Block Discovery.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Business Process Management
;
e-Business
;
Enterprise Engineering
;
Enterprise Information Systems
;
Knowledge Management and Information Sharing
;
Knowledge-Based Systems
;
Symbolic Systems
Abstract:
Process mining is a set of techniques helping enterprises to avoid process modeling which is a time-consuming and error prone task. Process mining includes three topics: process discovery, conformance checking, and enhancement (IEEE Task Force on Process Mining: Process Mining Manifesto, 2012). The principle of process discovery is to extract information from event logs to capture the business process as it is being executed. Several techniques in literature (α algorithm, α+ algorithm and others) can be applied to discover a process model from a workflow log. However, as the amount of information grows exponentially, the log files (input of a process discovery algorithm) get bigger. In fact, classical techniques, which inspect relation between each couple of tasks will have problem dealing with big data. To this end, we introduced in (Boushaba et al., 2013) a new approach aiming to extract a block of tasks from event logs. In this paper, we present a new algorithm, based on a matrix
representation, to detect a block of tasks. In addition, we develop an application to automate our technique.
(More)