A Survey on Concept Drift in Process Mining

Vecino, SatoDenise Maria; Cristiana, De FreitasSheila; Paul, BarddalJean; Emilio, ScalabrinEdson

doi:10.1145/3472752

Cited by 52 publications

(13 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Analyzing the impact of change factors, like the effect of process changes and dynamics in an ad‐hoc environment in detail is a must in this area. In this regard, proposed drift detection methods 23 and PARSs could be combined together. The reason is that PARSs require extensive knowledge of business processes for long‐term prediction considering multi‐classes based on a context, outcome, or process variant.…”

Section: Discussionmentioning

confidence: 99%

“…There are numerous review articles written in the PM: some have covered PM as a whole, 6,7 some have assessed the more detailed components like algorithms comparison [8][9][10] or a specific domain, for example, health care, [11][12][13][14] supply chain and industry, [15][16][17] and education. 18 There exist many review articles on process discovery, 19,20 conformance checking, 21,22 concept drift, 23 and predictive analytics [24][25][26][27][28] in PM. As to prescriptive analytics in PM, a literature review is conducted on prescriptive PM, 29 where merely the proposals in their implementation sense are assessed.…”

mentioning

confidence: 99%

See 1 more Smart Citation

A survey on recommendation in process mining

Eili

Rezaeenour

2022

Concurrency and Computation

View full text Add to dashboard Cite

Recommender systems have been widely applied in several domains to make informed decisions by recommending items that might be of interest. Considering recommendation during business process execution is also highly advantageous as the efficient suggestions about possible activities or resources can impact process performance. However, the deployment of the recommendation frameworks in process mining still needs more investigations to identify the current challenges to enable the practical application of research findings and ensure a large-scale adoption of this technique. Accordingly, a systematic review is conducted to provide a taxonomy of the published studies on process-aware recommender systems based on specified criteria, including the type and perspective of recommendation, a list of datasets and evaluation metrics used in the setting of PARS, implementation environments, and different algorithms used in PARS. In this regard, there are various insights extracted from this study: (i) Most studies in the business process analysis domain are of descriptive and predictive nature, (ii) recommendation in process mining is an emerging research area that is being evolving; the majority of proposals relate to 2015 and after that, and (iii) due to the lack of common evaluation protocol, datasets, and metrics, most studies are validated through experiments and prototyping, with less tendency to the practical implementation of a solution regarding real scenarios.

show abstract

Section: Discussionmentioning

confidence: 99%

mentioning

confidence: 99%

A survey on recommendation in process mining

Eili

Rezaeenour

2022

Concurrency and Computation

View full text Add to dashboard Cite

show abstract

“…Our generated high-level events are not coarser representations of the input event data, but rather events describing new emergent behavior caused by the combination of "low-level events"-none of which displays the behavior individually. Concept drift detection techniques [12] aim at identifying process changes that happen while the process is being analyzed. In our framework, significant process changes may become apparent as they cause the type of generated high-level events to also vary over time.…”

Section: Related Workmentioning

confidence: 99%

High-Level Event Mining: A Framework

Bakullari¹,

Aalst²

2022

Preprint

View full text Add to dashboard Cite

Process mining methods often analyze processes in terms of the individual end-to-end process runs. Process behavior, however, may materialize as a general state of many involved process components, which can not be captured by looking at the individual process instances. A more holistic state of the process can be determined by looking at the events that occur close in time and share common process capacities. In this work, we conceptualize such behavior using high-level events and propose a new framework for detecting and logging such high-level events. The output of our method is a new high-level event log, which collects all generated high-level events together with the newly assigned event attributes: activity, case, and timestamp. Existing process mining techniques can then be applied on the produced high-level event log to obtain further insights. Experiments on both simulated and real-life event data show that our method is able to automatically discover how system-level patterns such as high traffic and workload emerge, propagate and dissolve throughout the process.

show abstract

“…Organisations always try to adapt and evolve their business processes to handle different situations. Process mining techniques are expected to consider the concept drift challenge to allow process analysis in evolving businesses [ 7 ]. In particular, the process mining manifesto [ 8 ] identifies dealing with concept drifts in process mining among the main challenges in BPM.…”

Section: Introductionmentioning

confidence: 99%

PGraphD*: Methods for Drift Detection and Localisation Using Deep Learning Modelling of Business Processes

Hanga

Kovalchuk

Gaber

2022

Entropy

View full text Add to dashboard Cite

This paper presents a set of methods, jointly called PGraphD*, which includes two new methods (PGraphDD-QM and PGraphDD-SS) for drift detection and one new method (PGraphDL) for drift localisation in business processes. The methods are based on deep learning and graphs, with PGraphDD-QM and PGraphDD-SS employing a quality metric and a similarity score for detecting drifts, respectively. According to experimental results, PGraphDD-SS outperforms PGraphDD-QM in drift detection, achieving an accuracy score of 100% over the majority of synthetic logs and an accuracy score of 80% over a complex real-life log. Furthermore, PGraphDD-SS detects drifts with delays that are 59% shorter on average compared to the best performing state-of-the-art method.

show abstract

A Survey on Concept Drift in Process Mining

Cited by 52 publications

References 66 publications

A survey on recommendation in process mining

A survey on recommendation in process mining

High-Level Event Mining: A Framework

PGraphD*: Methods for Drift Detection and Localisation Using Deep Learning Modelling of Business Processes

Contact Info

Product

Resources

About