Translating Process Mining Results into Intelligible Business Information

Ceravolo, Paolo; Azzini, Antonia; Damiani, Ernesto; Lazoi, Mariangela; Marra, Manuela

doi:10.1145/2925995.2925997

Cited by 10 publications

(5 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is achieved by filtering the log based on well understood BRs. A preliminary study is described in Ceravolo et al [29]; it is a first analysis focus on a smaller dataset. This paper wants, through a case study method, to enlarge the collected evidence and better support and finalize how BRs can be used for a more valuable understanding of the process model and their feedback and impacts on the practice.…”

Section: Research Rationalementioning

confidence: 99%

See 1 more Smart Citation

Rules-based process mining to discover PLM system processes

Azzini¹,

Ceravolo

Corallo

et al. 2021

IJPLM

Self Cite

View full text Add to dashboard Cite

Process Mining (PM) is defined by a set of techniques used in Business Process Management that combines computational intelligence and data mining with process analysis and process modeling. The growing interest in PM is based on the ability to discover, monitoring, and improve the processes model. To reach this objective knowledge is extracted from the Event Logs generated by Information Systems, and quality metrics are used to evince the quality of the matching obtained when replaying a process model against the event log. The application of PM to logs extracted from PLM systems is an almost unexplored topic in this research area. Our study enhances the application in the field of PLM with the use of business rules to filter the log, verifying the BRs impact on PM metrics in order to minimize the divergences between modeled processes and executed one and to increase the resulting quality metrics. This helps the business user to identify a line of investigation for explaining occurring misbehavior and propose alleviation/improvement measures. Our approach is finally validated on data provided by an industrial company, by confirming the impact that controlling the business process characterizations via BR can decrease the gap between the expected modeled process and the executed one.

show abstract

Section: Research Rationalementioning

confidence: 99%

“…The complete set is available in Table 4. Respect to the previous preliminary study [29], having a larger dataset, we add the SBVR definition and we enlarge the set of rules.…”

Section: Rules Definitionmentioning

confidence: 99%

Rules-based process mining to discover PLM system processes

Azzini¹,

Ceravolo

Corallo

et al. 2021

IJPLM

Self Cite

View full text Add to dashboard Cite

show abstract

“…For example, they can isolate processes executed by a specific department, within a specific time-frame, or failing to achieve specific performance levels. As a result, these different variants can be comparatively assessed getting relevant insight into the enablers of specific behaviors [24,25]. Partially motivated by the same reasons, the authors of [26] proposed an ontology-based data access procedure to support the extraction of event logs from a legacy system, using a domain-specific conceptual model.…”

Section: Event-logs Standardsmentioning

confidence: 99%

Dynamic Access Control to Semantics-Aware Streamed Process Logs

et al. 2019

Self Cite

View full text Add to dashboard Cite

Business process logs are composed of event records generated, collected and analyzed at different locations, asynchronously and under the responsibility of different authorities. Their analysis is often delegated to auditors who have a mandate for monitoring processes and computing metrics but do not always have the rights to access the individual events used to compute them. A major challenge of this scenario is reconciling the requirements of privacy and access control with the need to continuously monitor and assess the business process. In this paper, we present a model, a language and a software toolkit for controlling access to process data where logs are made available as streams of RDF triples referring to some company-specific business ontology. Our approach is based on the novel idea of dynamic enforcement: we incrementally build dynamic filters for each process instance, based on the applicable access control policy and on the current prefix of the event stream. The implementation and performance validation of our solution is also presented.

show abstract

“…Process mining is a science that combines machine learning and data mining as well as processing modeling and analysis [12]. Process mining is a method used to generate business process models to find, monitor, and improve real processes by extracting knowledge from event logs, which are available in information systems derived from event log processes [3,11,13,21]. Process mining has three processes, including Discovery, Conformance Checking, and Enhancement [15].…”

Section: Introductionmentioning

confidence: 99%

Conformance Analysis of Student Activities to Evaluate Implementation of Outcome-Based Education in Early of Pandemic using Process Mining

et al. 2022

View full text Add to dashboard Cite

The learning management system has a core component of a system event log that contains data on activities carried out by students and lecturers in the system. Educational process mining is a field in educational data mining that is concerned with finding, analyzing, and improving the overall educational process based on information hidden in educational data sets and event logs. The learning process in student lectures through the learning management system will produce a process flow according to the event data. In one semester in the information technology-based study program, the subjects taken are data from programming and non-programming courses in the 5th semester of the information systems department, namely Data Warehouse and Business Intelligence (DWBI) and Enterprise Architecture (EA). The Data Warehouse and Business Intelligence course is chosen because the main role in the graduate profile is as a data engineer. While the Enterprise Architecture course is chosen because being an IT Consultant requires knowledge of EA. Each course has different measured learning outcomes and each course has a different pattern in obtaining learning outcomes. To get a pattern for each learning achievement, an analysis of learning patterns, Bloom’s taxonomy level, and CLO pass scores was carried out using process mining. Course Learning Outcomes (CLO) is a competency standard or minimum qualification criteria for graduates’ abilities which include attitudes, knowledge, and skills assigned to courses. The existence of a bloom level indicates the level of expected learning achievement, where the higher the bloom level, the higher the expected ability. The mining process is carried out using Disco and PROM 5.2. The modeling process uses a heuristic miner algorithm because of its ability to express the main behavior recorded in the event log well. Heuristic miner algorithm can also take into account the frequency of the relationship between activities in the log to determine causal dependencies. The results of this study indicate that there is a difference between those that pass the course learning outcomes and those that do not pass. The passed CLO is indicated by the realization value of passing the course exceeding the threshold of 85.50%, while the failed CLO is indicated by the realization value of course graduation that is less than the threshold. In addition, control-flow, the frequency of activities that are often carried out indicate the appropriate learning path and are carried out by students to achieve a minimal assessment of course learning outcomes. In the Enterprise Architecture course, the activity that has the highest frequency in CLO1 is Attempt Quiz, while in CLO6 is View Course. In the Data Warehouse and Business Intelligence course, the activity that has the highest frequency in CLO3 is View Course, while in CLO4 is Attempt Quiz. The initial activity of the learning pattern produced in the two courses begins always with the view course activity. The highest bloom level in the Data Warehouse and Business Intelligence course is C6 Creation, while in the Enterprise Architecture course is C5 Evaluation. Thus, it can be said that Data Warehouse and Business Intelligence courses have a higher level of difficulty than Enterprise Architecture. Previously, in the DWBI course there was one CLO that failed in its implementation. With this research, it is hoped that this research can have a positive impact on adding new insights regarding the use of event logs in the field of education, so implementation of outcome-based education can be used as a benchmark for student learning to succeed in the course which include attitudes, knowledge, and skills.

show abstract

Translating Process Mining Results into Intelligible Business Information

Cited by 10 publications

References 17 publications

Rules-based process mining to discover PLM system processes

Rules-based process mining to discover PLM system processes

Dynamic Access Control to Semantics-Aware Streamed Process Logs

Conformance Analysis of Student Activities to Evaluate Implementation of Outcome-Based Education in Early of Pandemic using Process Mining

Contact Info

Product

Resources

About