On the Use of Log-Based Model Checking, Clustering and Machine Learning for Process Behavior Prediction

Ezpeleta, J.; Fabra, Javier; Álvarez, Pedro

doi:10.1109/snams.2018.8554490

“…Park and Song (2020) followed a similar approach and mined an annotated transition system prior to their actual deep learning prediction. Ezpeleta et al (2018) applied linear-temporal-logic rules to the trace and used the resulting signatures as input for their FFNN to perform classification.…”

Section: Data Preprocessing and Feature Engineeringmentioning

confidence: 99%

“…On the chosen dataset, we selected the best variant, even if this variant did not dominate on all datasets. When per-class metrics were reported, as in Ezpeleta et al (2018), we display the arithmetic mean. Most authors executed the performance evaluation on a separate testing dataset to mitigate overfitting.…”

Section: Prediction Targetmentioning

confidence: 99%

“…Although most authors performed a binary classification, the underlying business problems are different and might be of varying difficulty. Nolle et al (2018) identified anomalistic traces, whereas Metzger and Neubauer (2018) predicted final events and Ezpeleta et al (2018) maped new traces to known When analysing scalar process performance measures, the results get even harder to compare, because the evaluation metrics are not scaled to a fixed interval. In this case, a meaningful comparison can only be made on identical data (see Fig.…”

Section: Prediction Targetmentioning

confidence: 99%

“…Like CNNs, these FFNNs suffer from a fixed input size forcing the authors to rescale prefixes of different lengths to a fixed size input matrix Nolle et al (2018). andPasquadibisceglie et al (2019) solved this restriction by padding all shorter prefixes to the maximum possible length Ezpeleta et al (2018). trained a separate model for multiple prefix lengths andMehdiyev et al (2020) used n-grams as inputs to model the temporal dependencies while preserving a fixed input size.…”

mentioning

confidence: 99%

See 2 more Smart Citations

A systematic literature review on state-of-the-art deep learning methods for process prediction

Neu

¹

,

Lahann

²

,

Fettke

³

2021

View full text Add to dashboard Cite

Process mining enables the reconstruction and evaluation of business processes based on digital traces in IT systems. An increasingly important technique in this context is process prediction. Given a sequence of events of an ongoing trace, process prediction allows forecasting upcoming events or performance measurements. In recent years, multiple process prediction approaches have been proposed, applying different data processing schemes and prediction algorithms. This study focuses on deep learning algorithms since they seem to outperform their machine learning alternatives consistently. Whilst having a common learning algorithm, they use different data preprocessing techniques, implement a variety of network topologies and focus on various goals such as outcome prediction, time prediction or control-flow prediction. Additionally, the set of log-data, evaluation metrics and baselines used by the authors diverge, making the results hard to compare. This paper attempts to synthesise the advantages and disadvantages of the procedural decisions in these approaches by conducting a systematic literature review.

show abstract

“…Although most authors performed a binary classification, the underlying business problems are different and might be of varying difficulty. Nolle et al (2018) identified anomalistic traces, whereas Metzger and Neubauer (2018) predicted final events and Ezpeleta et al (2018) maped new traces to known process clusters. Metzger and Neubauer only reported the MCC (0,626) and are therefore not listed in the table.…”

Section: Prediction Targetmentioning

confidence: 99%

A systematic literature review on state-of-the-art deep learning methods for process prediction

Neu,

Lahann,

Fettke

2021

Preprint

0

View full text Add to dashboard Cite

Process mining enables the reconstruction and evaluation of business processes based on digital traces in IT systems. An increasingly important technique in this context is process prediction. Given a sequence of events of an ongoing trace, process prediction allows forecasting upcoming events or performance measurements. In recent years, multiple process prediction approaches have been proposed, applying different data processing schemes and prediction algorithms. This study focuses on deep learning algorithms since they seem to outperform their machine learning alternatives consistently. Whilst having a common learning algorithm, they use different data preprocessing techniques, implement a variety of network topologies and focus on various goals such as outcome prediction, time prediction or control-flow prediction. Additionally, the set of log-data, evaluation metrics and baselines used by the authors diverge, making the results hard to compare. This paper attempts to synthesise the advantages and disadvantages of the procedural decisions in these approaches by conducting a systematic literature review.

show abstract