Exploiting CBOW and LSTM Models to Generate Trace Representation for Process Mining

Bui, Hong-Thuy; Sinh, Vu Trong; Nguyen, Hien-Hanh; Nguyen, Tri-Thanh; Ha, Quang-Thuy

doi:10.1007/978-981-15-3380-8_4

Cited by 3 publications

(3 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This vast research area called predictive business process monitoring, attracted several literature reviews (e.g., Neu et al (2021); Harane and Rathi (2020)). ML can also be used to optimise existing processes Fernandes et al (2019) or to get a compact representation of traces Bui et al (2019Bui et al ( , 2020. Recently, there has been interest in the interpretability of RNNs models, specifically in a process mining context Hanga et al (2020).…”

Section: Machine Learning For Process Monitoring and Miningmentioning

confidence: 99%

VaryMinions: leveraging RNNs to identify variants in variability-intensive systems’ logs

Fortz,

Temple,

Devroey

et al. 2024

Empir Software Eng

View full text Add to dashboard Cite

From business processes to course management, variability-intensive software systems (VIS) are now ubiquitous. One can configure these systems’ behaviour by activating options, e.g., to derive variants handling building permits across municipalities or implementing different functionalities (quizzes, forums) for a given course. These customisation facilities allow VIS to support distinct relevant customer requirements while taking advantage of reuse for common parts. Customisation thus allows realising both scope and scale economies. Behavioural differences amongst variants manifest themselves in event logs. To re-engineer this kind of system, one must know which variant(s) have produced which behaviour. Since variant information is barely present in logs, this paper supports this task by employing machine learning techniques to classify behaviours (event sequences) among variants. Specifically, we train Long Short Term Memory (LSTMs) and Gated Recurrent Units (GRUs) recurrent neural networks to relate event sequences with the variants they belong to on six different datasets issued from the configurable process and VIS domains. After having evaluated 20 different architectures of LSTM/GRU, our results demonstrate that it is possible to effectively learn the trace-to-variant mapping with high accuracy (at least $$80\%$$ 80 % and up to $$99\%$$ 99 % ) and at scale, i.e., identifying 50 variants using 5000+ traces for each variant.

show abstract

Section: Machine Learning For Process Monitoring and Miningmentioning

confidence: 99%

VaryMinions: leveraging RNNs to identify variants in variability-intensive systems’ logs

Fortz,

Temple,

Devroey

et al. 2024

Empir Software Eng

View full text Add to dashboard Cite

show abstract

“…De Koninck et al [4] transferred the idea of Word2vec [8] and Doc2vec [7] to process data. An LSTM and CBOW-based approach was introduced by Bui et al [2]. A supervised representation learning approach based on conditional random fields for event abstraction was introduced by Tax et al [14].…”

Section: Related Workmentioning

confidence: 99%

“…In our parameter optimization strategy, we first optimize the vector size. We vary the vector size of the hidden and the embedding layer (2,3,4,8,16,32,64,128,256), and the number of epochs (10,25,50). Next, we optimize the window size of the embedding which determines how many activities before and after the current activity are considered.…”

Section: Real-life Event Logs: Trace Clusteringmentioning

confidence: 99%

Case2vec: Advances in Representation Learning for Business Processes

Luettgen

Seeliger

Nolle

et al. 2021

Lecture Notes in Business Information Processing

View full text Add to dashboard Cite

The execution of a business process is often determined by the surrounding context, e.g., department, product, or other attributes an event provides. Process discovery mainly focuses on the executed activities, although the context of a case may be needed to accurately represent a process instance, e.g., for clustering, prediction, or anomaly detection. Hence, in this paper, we present a representation learning technique (Case2vec) using word embeddings for business process data to better encode process instances. Our work extends Trace2vec and incorporates an additional semantic level by using not only the activity name but also the attributes and thereby incorporating the context. We evaluate our approach in the context of trace clustering. Additionally, we show that Case2vec can be used to abstract events which are semantically similar but syntactically different. We also show that word embeddings allow for interpretability when employing vector space arithmetic.

show abstract

Third Order Polynomial Derived Sentence Embedding

Sithole¹,

Marshall²

2021

2021 International Conference on Electrical, Computer and Energy Technologies (ICECET)

View full text Add to dashboard Cite

Exploiting CBOW and LSTM Models to Generate Trace Representation for Process Mining

Cited by 3 publications

References 7 publications

VaryMinions: leveraging RNNs to identify variants in variability-intensive systems’ logs

VaryMinions: leveraging RNNs to identify variants in variability-intensive systems’ logs

Case2vec: Advances in Representation Learning for Business Processes

Third Order Polynomial Derived Sentence Embedding

Contact Info

Product

Resources

About