To assess post-run time efficiency of business processes, details about these processes are collected and then, stored in event logs upon which mining is carried out. Commonly adopted in the literature, this mining suffers from several limitations such as lack of details about the policies that either authorized or denied the consumption of resources by business processes. This paper addresses these limitations through policy logs complementing event logs during process mining. A policy log captures which policies were triggered, by whom, and what the outcome was. In term of implementation, the use of a BPI Challenge 2017 dataset shows that policy-only replay reveals violations that are invisible to control-flow analysis. And, a controlled synthetic evaluation with injected, labeled violations enables precision, recall, and PR-AUC reporting for policy-only conformance checking.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Complementing Event Log with Policy Log for Business Process Mining

  • Zakaria Maamar,
  • Amel Benna,
  • Abderrahmane Maaradji

摘要

To assess post-run time efficiency of business processes, details about these processes are collected and then, stored in event logs upon which mining is carried out. Commonly adopted in the literature, this mining suffers from several limitations such as lack of details about the policies that either authorized or denied the consumption of resources by business processes. This paper addresses these limitations through policy logs complementing event logs during process mining. A policy log captures which policies were triggered, by whom, and what the outcome was. In term of implementation, the use of a BPI Challenge 2017 dataset shows that policy-only replay reveals violations that are invisible to control-flow analysis. And, a controlled synthetic evaluation with injected, labeled violations enables precision, recall, and PR-AUC reporting for policy-only conformance checking.