Process Discovery from Low-Level Event Logs

Fazzinga, Bettina; Flesca, Sergio; Furfaro, Filippo; Pontieri, Luigi

doi:10.1007/978-3-319-91563-0_16

Cited by 11 publications

(8 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Each possible interpretation of the fine-granular trace, in terms of coarse-granular events, is computed with an associated probability distribution. In Fazzinga et al (2018b), the authors use a hidden Markov model that models the generation of low-level events. The frequency of precedence relations between high-level activities are inferred.…”

Section: Supervised Techniquesmentioning

confidence: 99%

“…For example, in Tax et al (2018), the authors propose to use linear-chain conditional random fields (CRFs) to represent sequences of sensor data. Likewise, in Alharbi et al (2018) and Fazzinga et al (2018b), the authors propose to use hidden Markov chains as an internal representation.…”

Section: Fine-granular Event Interleavingmentioning

confidence: 99%

“…The estimation is based on the time differences between start and complete lifecycle steps for an activity. In Fazzinga et al (2018b), the authors propose to use a hidden Markov model that models the generation of fine-granular events, which is used to infer the frequency of precedence relations, rather than coarse-granular events.…”

Section: Probabilistic Nature Of Outcomementioning

confidence: 99%

See 2 more Smart Citations

Event abstraction in process mining: literature review and taxonomy

et al. 2020

View full text Add to dashboard Cite

The execution of processes in companies generates traces of event data, stored in the underlying information system(s), capturing the actual execution of the process. Analyzing event data, i.e., the focus of process mining, yields a detailed understanding of the process, e.g., we are able to discover the control flow of the process and detect compliance and performance issues. Most process mining techniques assume that the event data are of the same and/or appropriate level of granularity. However, in practice, the data are extracted from different systems, e.g., systems for customer relationship management, Enterprise Resource Planning, etc., record the events at different granularity levels. Hence, pre-processing techniques that allow us to abstract event data into the right level of granularity are vital for the successful application of process mining. In this paper, we present a literature study, in which we assess the state-of-the-art in the application of such event abstraction techniques in the field of process mining. The survey is accompanied by a taxonomy of the existing approaches, which we exploit to highlight interesting novel directions.

show abstract

Section: Supervised Techniquesmentioning

confidence: 99%

Section: Fine-granular Event Interleavingmentioning

confidence: 99%

Section: Probabilistic Nature Of Outcomementioning

confidence: 99%

See 1 more Smart Citation

Event abstraction in process mining: literature review and taxonomy

et al. 2020

View full text Add to dashboard Cite

show abstract

“…This is the reason why our approach uses a lossless representation of event logs (DAFSAs) to discover candidate automatable routines, as opposed to an automatically discovered process model. Recent related work in APD deals with the problem of discovering process models from low-level event logs [8]. In this context, a low-level event log is one where each event refers to a step of a task in a process, e.g.…”

Section: Related Workmentioning

confidence: 99%

“…a task "Contact customer" is captured via several events corresponding to steps "Retrieve the customer's contact details from CRM system", "call customer", etc. The goal of [8] and related studies is to group low-level events into coarser-grained ones in order to discover conceptual models, as opposed to automatable routines as we do.…”

Section: Related Workmentioning

confidence: 99%

Discovering Automatable Routines from User Interaction Logs

Bosco

Augusto

Dumas

et al. 2019

Lecture Notes in Business Information Processing

View full text Add to dashboard Cite

The complexity and rigidity of legacy applications in large organizations engender situations where workers need to perform repetitive routines to transfer data from one application to another via their user interfaces, e.g. moving data from a spreadsheet to a Web application or vice-versa. Discovering and automating such routines can help to eliminate tedious work, reduce cycle times, and improve data quality. Advances in Robotic Process Automation (RPA) technology make it possible to automate such routines, but not to discover them in the first place. This paper presents a method to analyse user interactions in order to discover routines that are fully deterministic and thus amenable to automation. The proposed method identifies sequences of actions that are always triggered when a given activation condition holds and such that the parameters of each action can be deterministically derived from data produced by previous actions. To this end, the method combines a technique for compressing a set of sequences into an acyclic automaton, with techniques for rule mining and for discovering data transformations. An initial evaluation shows that the method can discover automatable routines from user interaction logs with acceptable execution times, particularly when there are one-to-one correspondences between parameters of an action and those of previous actions, which is the case of copy-pasting routines.

show abstract

Extraction, correlation, and abstraction of event data for process mining

Diba

Batoulis

Weidlich

et al. 2019

WIREs Data Min & Knowl

View full text Add to dashboard Cite

Process mining provides a rich set of techniques to discover valuable knowledge of business processes based on data that was recorded in different types of information systems. It enables analysis of end-to-end processes to facilitate process reengineering and process improvement. Process mining techniques rely on the availability of data in the form of event logs. In order to enable process mining in diverse environments, the recorded data need to be located and transformed to event logs. The journey from raw data to event logs suitable for process mining can be addressed by a variety of methods and techniques, which are the focus of this article. In particular, techniques proposed in the literature to support the creation of event logs from raw data are reviewed and classified. This includes techniques for identification and extraction of the required event data from diverse sources as well as their correlation and abstraction.

show abstract

Process Discovery from Low-Level Event Logs

Cited by 11 publications

References 21 publications

Event abstraction in process mining: literature review and taxonomy

Event abstraction in process mining: literature review and taxonomy

Discovering Automatable Routines from User Interaction Logs

Extraction, correlation, and abstraction of event data for process mining

Contact Info

Product

Resources

About