A Comprehensive and Automated Approach to Intelligent Business Processes Execution Analysis
Malu Castellanos,Fabio Casati,Umeshwar Dayal,Ming-Chien Shan
DOI: https://doi.org/10.1023/B:DAPD.0000031635.88567.65
IF: 0.974
2004-01-01
Distributed and Parallel Databases
Abstract:Business process management tools have traditionally focused on supporting the modeling and automation of business processes, with the aim of enabling faster and more cost-effective process executions. As more and more processes become automated, customers become increasingly interested in managing process executions. Specifically, there is a desire for getting more visibility into process executions, to be able to quickly spot problems and areas for improvements. The idea is that, by being able to assess the process execution quality, it is possible to take actions to improve and optimize process execution, thereby leading to processes that have higher quality and lower costs. All this is possible today, but involves the execution of specialized data mining projects that typically last months, costs hundreds of thousands of dollars, and only provide a specialized, narrow solution whose applicability is often relatively short in time, due to the ever changing business and IT environments. Still, the need is such that companies undertake these efforts. To address these needs, this paper presents a set of concepts and architectures that lay the foundation for providing users with intelligent analysis and predictions about business process executions. For example, the tools are able to provide users with information about why the quality of a process execution is low, what will be the outcome of a certain process, or how many processes will be started next week. This information is crucial to gain visibility into the processes, understand or foresee problems and areas of optimization, and quickly identify solutions. Intelligent analysis and predictions are achieved by applying data mining techniques to process execution data. In contrast to traditional approaches, where lengthy projects, considerable efforts, and specialized skills in both business processes and data mining are needed to achieve these objectives, we aim at automating the entire data mining process lifecycle, so that intelligent functionality can be provided by the system while requiring little or no user input. The ambitious end goal of the work presented in this paper is that of laying the foundation for a framework and tool that is capable of providing analysts with key intelligence information about process execution, affecting crucial IT and business decisions, almost literally at the click of a button.