A clustering based incremental faulty-rate estimation algorithm for business process monitoring

Abstract

This paper proposes a novel approach to real-time business process monitoring using a newly proposed clustering based incremental faulty-rate estimation algorithm. In our approach, faulty-rate is defined as a probability that an ongoing process might be ended in fault, which can be estimated based on observed attributes and possible outcomes at each monitoring phase. In the proposed estimation algorithm, unobserved attributes are substituted by historical distributions so that the faulty-rate can be derived in a distribution function. Finally, how the faulty-rate estimation can be applied to the real-time process monitoring is illustrated with an example scenario. ICIC International © 2011 ISSN 1881-803X.