Discriminative Rule Learning for Outcome-Guided Process Model Discovery

TL;DR

The paper tackles outcome-aware process discovery by learning discriminative control-flow patterns that separate desirable $L^+$ from undesirable $L^-$ traces. It combines declarative constraint encoding, ensemble tree feature extraction, and sparse regression to identify representative discriminative rules, then clusters traces by these rules to perform cluster-specific process discovery. The approach yields focused, interpretable variants that better reveal drivers of desirable and undesirable executions, with empirical evidence on real logs showing improved discriminative metrics over baselines. The work provides an open-source tool and a framework for outcome-guided conformance analysis, highlighting practical impact for conformance checking and performance analysis with outcome-aware models.

Abstract

Event logs extracted from information systems offer a rich foundation for understanding and improving business processes. In many real-world applications, it is possible to distinguish between desirable and undesirable process executions, where desirable traces reflect efficient or compliant behavior, and undesirable ones may involve inefficiencies, rule violations, delays, or resource waste. This distinction presents an opportunity to guide process discovery in a more outcome-aware manner. Discovering a single process model without considering outcomes can yield representations poorly suited for conformance checking and performance analysis, as they fail to capture critical behavioral differences. Moreover, prioritizing one behavior over the other may obscure structural distinctions vital for understanding process outcomes. By learning interpretable discriminative rules over control-flow features, we group traces with similar desirability profiles and apply process discovery separately within each group. This results in focused and interpretable models that reveal the drivers of both desirable and undesirable executions. The approach is implemented as a publicly available tool and it is evaluated on multiple real-life event logs, demonstrating its effectiveness in isolating and visualizing critical process patterns.

Figures (7)

• Figure 1: Trace variants and frequencies in $L^+$ and $L^-$.
• Figure 2: Process models for desirable and undesirable traces.
• Figure 3: Decision tree classifying traces into desirable or undesirable.
• Figure 4: An overview of our proposed framework.
• Figure 5: Jaccard distance heatmaps between extracted rules for each event log. Lighter blue indicates greater similarity. Rule support (orange) and importance (green-red) are shown on the left. Dendrograms guide the identification of trace clusters.

Theorems & Definitions (10)

• definition thmcounterdefinition: Event Log
• definition thmcounterdefinition: Case Label
• definition thmcounterdefinition: Declarative Constraints
• definition thmcounterdefinition: Declarative Constraints Evaluation
• definition thmcounterdefinition: Feature Space
• definition thmcounterdefinition: Rule Extraction
• definition thmcounterdefinition: Rule-based Encoding
• definition thmcounterdefinition: Regression Classifier
• definition thmcounterdefinition: Jaccard Distance between Rules
• definition thmcounterdefinition: Discriminative Evaluation Metrics