Enhancing similarity measures and context provision for the intelligent monitoring of business processes in CBR-WIMS

Kapetanakis, Stelios, Petridis, Miltos, Ma, Jixin, Knight, Brian and Bacon, Liz (2011) Enhancing similarity measures and context provision for the intelligent monitoring of business processes in CBR-WIMS. In: 19th International Conference on Case Based Reasoning: Workshop 1 - Process-oriented Case-Based Reasoning (PO-CBR), 12-15 September 2011, Greenwich, London. (Submitted)

PDF - Submitted Version Restricted to Repository staff only Download (1986kB)

Abstract

Business processes have a key role in operating, controlling, and managing large modern organizations. Managing business processes presents a challenge related to the temporal complexity, uncertainty and large volume of data generated. This paper presents new developments and evaluation of an enhanced approach for the intelligent monitoring of business processes using Case-Based Reasoning in the CBR-WIMS platform. A short overview of the CBR-WIMS approach, based on the representation of business process cases as graphs, comprising process events and their temporal relationships is presented. The enhanced similarity measure based on the Maximum Common Sub-graph is presented and an evaluation of its effectiveness and efficiency is shown and discussed. The evaluation uses historical data from a real business process. The paper also discusses the use of a clustering technique in CBR-WIMS. This allows the semi-automatic tagging of cases and so provides enhanced explanation and context to users, increasing the confidence and usability of retrieved solutions and advice. A set of experiments are presented and discussed showing the added value that this enhancement brings to the intelligent monitoring and management of real business processes in an organisation.

Actions (login required)