CONDA-PM -- A Systematic Review and Framework for Concept Drift Analysis in Process Mining

Business processes evolve over time to adapt to changing business environments. This requires continuous monitoring of business processes to gain insights into whether they conform to the intended design or deviate from it. The situation when a business process changes while being analysed is denoted as Concept Drift. Its analysis is concerned with studying how a business process changes, in terms of detecting and localising changes and studying the effects of the latter. Concept drift analysis is crucial to enable early detection and management of changes, that is, whether to promote a change to become part of an improved process, or to reject the change and make decisions to mitigate its effects. Despite its importance, there exists no comprehensive framework for analysing concept drift types, affected process perspectives, and granularity levels of a business process. This article proposes the CONcept Drift Analysis in Process Mining (CONDA-PM) framework describing phases and requirements of a concept drift analysis approach. CONDA-PM was derived from a Systematic Literature Review (SLR) of current approaches analysing concept drift. We apply the CONDA-PM framework on current approaches to concept drift analysis and evaluate their maturity. Applying CONDA-PM framework highlights areas where research is needed to complement existing efforts.

💡 Research Summary

The paper addresses the problem of concept drift in business processes—situations where the underlying process changes while it is being analyzed. Recognizing the lack of a unified framework that links drift types, affected process perspectives, and granularity levels, the authors conduct a systematic literature review (SLR) of 68 recent studies on drift detection, localisation, impact analysis, and mitigation. From this review they derive CONDA‑PM (Concept Drift Analysis in Process Mining), a comprehensive framework that structures drift analysis into four sequential phases: (1) Detection, employing statistical tests, machine learning, and time‑series models; (2) Localisation, identifying the exact process elements (activities, transitions, resources) that have changed using transition matrices, dynamic sub‑graph matching, and pattern mining; (3) Impact Analysis, quantifying the effect of the drift on key performance indicators such as cycle time, cost, quality, and resource utilisation through causal inference, Bayesian networks, and simulation‑based scenario analysis; and (4) Decision‑Making, selecting whether to adopt the drift as an improvement or to mitigate it using cost‑benefit analysis, multi‑objective optimisation, and decision trees.

CONDA‑PM explicitly maps each phase to four process perspectives—control‑flow, data, resource, and time—and to three granularity levels—trace, event, and segment. For example, a control‑flow drift may be detected at the trace level as a shift in overall model conformance, while at the event level it may appear as a new activity ordering. Data drifts involve changes in attribute distributions, resource drifts capture role swaps or staffing changes, and time drifts reflect variations in cycle time or seasonality. By integrating these dimensions, the framework overcomes the common limitation of prior work that focuses almost exclusively on control‑flow at a single granularity.

Applying CONDA‑PM to the surveyed literature, the authors construct a maturity matrix. They find that detection and localisation techniques are relatively mature, with many studies offering robust statistical or deep‑learning based solutions. In contrast, impact analysis and decision‑making remain under‑developed; few papers provide quantitative causal models or real‑time mitigation mechanisms, and economic evaluation of drift management is largely absent. The paper therefore highlights several research gaps: (i) the need for integrated multi‑perspective, multi‑granularity analysis methods; (ii) advanced causal inference models that can link drift to business outcomes; (iii) automated, real‑time decision support systems that incorporate cost‑benefit reasoning; and (iv) open‑source toolkits that implement the full CONDA‑PM pipeline.

In conclusion, CONDA‑PM offers a systematic, extensible roadmap for researchers and practitioners to move beyond simple drift detection toward a holistic understanding of how process changes affect performance and how organizations can strategically respond. By providing clear phase definitions, perspective‑granularity mappings, and a maturity assessment of existing approaches, the framework sets the stage for more rigorous, value‑driven process mining research and practice.