Predictive monitoring of temporally-aggregated performance indicators of business processes against low-level streaming events

Highlights：

• We focus on the context of complex event-based information system monitoring.

• We provide a framework for supporting predictive monitoring of process performance indicators.

• We face the problem of predicting whether (the process instances in) each current time window will violate an aggregate performance constraint, at different checkpoints inside the window.

• Our prediction task is accomplished by inducing two kinds of models, which both take advantage of aggregate information on the state of the environment where the process is being executed: (1) a clustering-based model for estimating the outcome of each unfinished process instance, (2) ad-hoc time-series models for estimating global statistics on the outcome of future process instances that are expected to start by the end of the window.

• We propose an event-based architecture for an innovative kind of monitoring infrastructure that integrates the proposed prediction approach.

• We provide tests on real-life event data that confirmed the validity of the approach, in terms of both prediction accuracy and scalability.

摘要

•We focus on the context of complex event-based information system monitoring.•We provide a framework for supporting predictive monitoring of process performance indicators.•We face the problem of predicting whether (the process instances in) each current time window will violate an aggregate performance constraint, at different checkpoints inside the window.•Our prediction task is accomplished by inducing two kinds of models, which both take advantage of aggregate information on the state of the environment where the process is being executed: (1) a clustering-based model for estimating the outcome of each unfinished process instance, (2) ad-hoc time-series models for estimating global statistics on the outcome of future process instances that are expected to start by the end of the window.•We propose an event-based architecture for an innovative kind of monitoring infrastructure that integrates the proposed prediction approach.•We provide tests on real-life event data that confirmed the validity of the approach, in terms of both prediction accuracy and scalability.