S4BP: An Approach for Assessing Business Process Stability

Hajer Ben Haj Ayech

1,4 a

, Ricardo Martinho

and Sonia Ayachi Ghannouchi

3,4 c

Higher Institute of Computer Science and Communication Technologies of Sousse, University of Sousse, Tunisia

INESCC – DL, ESTG, Polytechnic University of Leiria, Portugal

Higher Institute of Management of Sousse, University of Sousse, Tunisia

RIADI Laboratory, University of Manouba, Tunisia

Keywords: Business Process Stability, Process Mining, Metrics, Evaluation, Prediction, Approach.

Abstract: Achieving business process (BP) stability is a fundamental objective for organizations, pursued for a variety

of reasons including consistency in operations and product/service delivery, reduced costs and rework, and

clear metrics for process improvement. Nevertheless, the subject has received little attention in research, from

vague definitions to mingled concepts involving BP flexibility and changes. This paper addresses the stability

of BP in the context of Business Process Management (BPM). Specifically, it proposes a clearer definition of

BP stability, as well as a step-by-step Stability for Business Processes approach (S4BP) based on Process

Mining techniques to evaluate and predict stability for a certain BP. The proposed approach is demonstrated

through a software implementation in the form of a ProM plugin, and validated using a case study with public

datasets from the Business Process Improvement (BPI) Challenge.

1 INTRODUCTION

Business Process Management (BPM) serves as a

valuable approach for addressing organizational and

strategic challenges by promoting both stability and

flexibility in business process (BP) models (Cognini

et al, 2016). Stability control is critical in BP, as it

enhances customer satisfaction. By maintaining stable

processes, organizations can ensure timely delivery of

products or services, which significantly contributes

to customer satisfaction and loyalty. Moreover,

assessing process stability is crucial for maintaining

high operational quality (Willis et al, 2018). Stable

processes provide a robust foundation for informed

decision-making and support continuous

improvements for smoother business operations.

It is beneficial for a process to be both innovative

and capable of adapting to changes (Ben Haj Ayech et

al, 2021). However, it is also necessary to have stable

processes that resist modifications across different

versions, as this is essential to ensure the reliability

and consistency in the long term (Kelly, 2006).

https://orcid.org/0009-0003-9939-3712

https://orcid.org/0000-0003-1157-7510

https://orcid.org/0000-0001-9583-9797

A stable environment thus promotes rational

decision-making through comparative solution

evaluation, rather than being driven by pressing

deadlines. This strengthens the organization’s ability

to optimize its processes, improve execution, and

adapt its choices in response to changing conditions

and emerging opportunities.

BP models exhibit considerable dynamism and

often require modifications (Ben Haj Ayech et al,

2021). According to Baumgraß et al. (2014), process

data can evolve over time, making its correlation with

processes particularly complex. This dynamic nature

of process data can have significant implications for

process stability. Therefore, there is a need to assess

the stability of business processes by analyzing

variations over specific periods of time, to ensure

their reliability and consistency.

Despite exiting several research works that offer

various definitions on the concept of stable processes

in different domains, BP stability is, to our

knowledge, yet to be addressed in the specific domain

of BPM. This work presents our contribution to the

theme, beginning by the proposal of a definition of

stability specific to BP, supported by a systematic

208

Ayech, H. B. H., Martinho, R. and Ghannouchi, S. A.

S4BP: An Approach for Assessing Business Process Stability.

DOI: 10.5220/0013427600003928

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 20th Inter national Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2025), pages 208-216

ISBN: 978-989-758-742-9; ISSN: 2184-4895

approach and an implementation of a plugin in the

ProM tool for validation, including the use of Process

Mining techniques over a public dataset of BP.

The paper is structured as follows: Section 2

provides background on stability, BP and Process

Mining techniques, highly involved on our proposed

approach. In Section 3, we address related work, and

in Section 4 we propose our S4BP approach. Section

5 presents the implementation and validation of the

approach on ProM, and finally, Section 6 concludes

the paper and presents future work.

2 BACKGROUND

The concept of stability is generally implicit,

acknowledged, but rarely formally defined. Few

precise definitions of stability are available, and these

are mostly associated with software development

processes. For instance, the work of (Yau et al, 1980)

introduced a definition of code stability, followed

later by a definition of design stability (Yau et al,

1985). Another definition is proposed by (Kelly,

2006) for the stability of a design characteristic,

suggesting that if the value of the metric associated

with that characteristic exhibits slight variations

between two or more versions of the software, then

that characteristic can be considered stable.

According to this study, limited variation indicates

that, despite the changes made, the design retains

fundamental elements that remain constant.

Despite the importance of BP stability and its

impact on various organizational factors, much of the

existing literature does not address the measurement

of stability in BP or the prediction of their subsequent

changes. On the contrary, the concept of BP

flexibility was earlier defined for instance, in (Daoudi

et al, 2005), (Pesic et al, 2006), (Schonenberg et al,

2008

) and (Regev et al, 2006) and is widely used in

literature, including well defined ways to measure it

(Mejri et al, .2018), (Mejri et al, 2024).

Nevertheless, process stability is a significant

concept in Business Process Management (BPM) that

can also be defined and measured using various

quality metrics, such as precision, complexity, and

outcome prediction (Jongchan, 2021). For this

purpose, Process Mining includes data science

techniques that aim to discover, monitor, and improve

actual BP by extracting valuable metrics and insights

from event logs (Van der Aalst et al, 2021). The

objective of process discovery is to automatically

identify the schema of a process from an event log

(Van der Aalst et al, 2011). Although the majority of

BP undergo dynamic evolution over time, often in

response to internal and external factors, current

process mining approaches often assume processes

are in a stable state. Consequently, an increasing

number of algorithms have been developed to

compare different variants of the same process

(Hompes et al, 2015) (Luengo et al, 2012). The

primary aim of process discovery involves the

automated extraction of a process schema from event

logs (Van der Aalst et al, 2011). As a result, a growing

array of algorithms has been developed to analyze

different variants of the same process or to identify

shifts in processes over time (Lavanya et al, 2015).

These algorithms play a vital role in recognizing and

comprehending alterations in BP, thus enabling

organizations to seize new opportunities and ensure

ongoing improvement.

Moreover, Process Mining can be employed

iteratively, facilitating the creation of more

comprehensive data records, including actor names,

notes, dates, and the curriculum elements such as

required work, essential concepts to be grasped, and

the instructor's primary goals. This iterative approach

enables swift identification and resolution of

encountered issues for future enhancement. Process

mining techniques aim to convert data collected

during process execution into actionable information

and knowledge (Bergaoui et al, 2024).

3 RELATED WORK

This section provides an overview of existing

research related to BP stability. We begin by

discussing BP discovery techniques, which can serve

as a foundation for any stability assessment.

Understanding how processes are extracted and

represented is crucial for subsequent analysis. Next,

we explore various metrics used to evaluate BP

stability, highlighting key indicators that help

quantify stability and its influencing factors. Finally,

we identify existing research gaps in BP stability

studies, emphasizing the need for further

investigation and positioning our contribution within

this context.

3.1 BP Discovery

One of the most studied Process Mining techniques is

the automated discovery of processes (Adriano et al,

2018). These techniques take an event log as input

and generates a BP model that captures the control-

flow relationships among tasks observed or inferred

from the event log. The model produced must meet

several criteria: it should be able to generate each

S4BP: An Approach for Assessing Business Process Stability

209

trace present in the log, create traces similar to those

in the event log, and produce traces that are not in the

log but are identical or similar to the traces of the

process that generated the log.

Several approaches have been proposed in the

literature to identify changes made to BP models. The

work of (Günther et al, 2008) integrates Process

Mining with adaptive process management, utilizing

log files from Process Mining to enhance adaptive

management systems. Their research emphasizes a

flexibility metric that facilitates modifications and

changes in dynamic BP models during execution. By

employing Process Mining as an analytical tool, they

offer insights into when and why process changes

become necessary, thereby improving support for

flexible processes. (Berti, 2016) further develops

Process Mining techniques to enhance the prediction

and detection of dynamic changes in BP using various

algorithms, statistical tests, and probabilistic

approaches. Another significant contribution is

BPMN-CM (Business Process Model and Notation

Change Management), introduced by (Kherbouche,

2013), which assists in managing the evolution of BP

models by analyzing the impact of changes to ensure

model consistency after each modification (Ben Haj

Ayech et al, 2021). Moreover, (Maaradji et al, 2017)

focus on extracting information regarding change

techniques in systems by analyzing collected data to

compute relevant timestamp differences. Their

method, which detects progressive drifts, represents a

family of techniques aimed at identifying changes in

BP. Their empirical evaluation demonstrates that this

method achieves higher accuracy and shorter

detection times in identifying typical change patterns

compared to existing methods. Additionally,

(Alejandro et al, 2018) utilized interaction data from

101 university students, mining 21 629 events to

assess the models produced by different algorithms in

terms of fitness, precision, generalization, and

simplicity metrics. They compared results from

algorithms such as Heuristic Miner, Evolutionary

Tree Miner, Alpha Miner, and Inductive Miner,

finding that the Inductive Miner algorithm yielded the

best performance overall, particularly when various

metrics were weighted.

(Carlos, 2022) compared Process Mining tools

and algorithms, noting that Alpha Miner, as the initial

algorithm for process discovery, generates a Petri net

model by first identifying existing traces, analyzing

the sequence of activities, and creating a relationship

matrix. The study also highlights Heuristic Miner,

which constructs the process map by considering the

frequency of events rather than solely the sequence,

focusing on the most frequent paths while

disregarding those that appear less often.

Although these related works mention important

BP flexibility and changeability metrics and

approaches, we have observed a significant lack of

research focusing on the stability of BP, despite their

crucial importance for the operational efficiency of

organizations. This work aims to establish a

conceptual framework that will allow us to contribute

to a better understanding of the interactions involving

the definition, assessment and prediction BP stability.

3.2 BP Metrics Related to Stability

This section discusses metrics that can be used to

compute BP stability. According to (Adriano et al,

2018), the Process Mining-related fitness metric

measures a BP model's capacity to reproduce the

behaviors represented in an event log, where a score

of 1 indicates complete reproduction of all traces.

Precision assesses the model's ability to generate only

the behaviors found in the log, with a score of 1

indicating that all traces produced by the model are

contained within the log.

The frequency and extent of modifications made

to process models over time are captured by the

number of changes. These changes can stem from

various factors, such as evolving business needs, new

regulations, or efforts to improve performance criteria

(Philip et al, 2011). The number of variants in BPM

serves as a metric for assessing process model

variability, which is contingent upon the specific

needs and circumstances of the organization (Fredrik

et al, 2012). Model redundancy in BPM is another

crucial metric, referring to the presence of duplicate

or unnecessary elements within BPM models, which

may lead to confusion, inefficiencies, and errors (Fei

et al, 2021).

Generally, these metrics are evaluated for a

certain period of time within process event logs,

ranging between BP perspectives such as control-

flow (the sequencing, frequency and timing of

activities), resources (e.g., teams allocated to

activities) or data (documents’ states along process

execution). Nevertheless, by themselves these

metrics are poor to assess BP stability, since it usually

implies calculation logic and evolution over time.

Additionally, prediction techniques can be further

applied around this calculation logic, as for example

using linear regression predict various events, which

not only enables effective management of product

quality but also allows for the analysis of a wide range

of data (Gezani et al, 2015).

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

210

3.3 BP Stability Research Gap

From our point of view, organizations can pursue

stability in their BP for several reasons. A stable

process guarantees consistent and predictable

outcomes, which are essential for fulfilling customer

expectations and sustaining a strong reputation.

Furthermore, stable processes facilitate smooth

operations, minimizing the need for frequent

adjustments, thereby increasing efficiency and

productivity. Additionally, a stable process

contributes to cost reduction by decreasing material

waste and unexpected downtimes, resulting in cost

savings and more environmentally friendly

operations. Timely delivery of products or services

from a stable process enhances customer satisfaction

and loyalty, while also enabling the achievement of

desired outcomes and meeting customer

specifications. Moreover, stable processes provide a

solid foundation for predictability, informed

decision-making and optimization of business

operations, allowing organizations to swiftly adapt to

changing market conditions and scale their operations

accordingly.

To the best of our knowledge, this concept of BP

stability and the way it can be achieved remains a

rather unexplored research theme.

4 THE S4BP APPROACH

In this section, we present the S4BP approach, which

aims to assess and ensure the stability of BP. First, we

define the concept of business process stability,

identifying its key characteristics and its importance

in the context of our study. Then, we detail our

methodological approach, explaining the steps

followed, the principles adopted, and the tools used to

evaluate and predict the BP stability.

4.1 BP Stability Definition

In this paper, we consider a business process to be

stable when it evolves in a controlled and predictable

manner, based on the analyzed perspectives, applied

metrics, and their trends over time. We define BP

stability as the ability of a business process to

maintain structural consistency and operational

predictability in the face of changes and evolutions.

In addition, we introduce a formal definition,

encompassing three main process key components:

perspectives, metrics and trend analysis. Perspectives

cover the process control-flow, data, and resource

dimensions. Metrics include calculated values over

the data of these perspectives. Examples include

similarity metrics, number of variants, change

frequency, and process fitness. Trend analysis

consider the evolution of these metrics over different

periods of time.

Building upon these foundational concepts, we

propose a formulation where PS represents process

stability, defined as a function of various factors,

including distinct process perspectives, metrics

calculation logic, and trend analysis. We can then

formulate the process stability as follows:

𝑃𝑆 = f



PP, T, M, C



(1)

Where:

 PP refers to the process perspectives, including

control-flow (CF), data (D), and resources (R),

defined as:

PP= {CF, D, R} (2)

 M denotes the set of metrics used for assessing

stability, defined as:

M =



metric1, metric2, … , metric𝑛



(3)

 T is trend analysis, representing time intervals

over which metrics are evaluated, expressed as:

T =



t1 , t2, … , tn



(4)

 C represents the stability calculation logic,

involving an aggregation function over the

specified trend analysis time periods t.

Having this formulation, we can now proceed

with the proposal of an approach to apply it

concretely.

4.2 Approach

Our proposed S4BP approach foresees the application

of the previous definition into five key phases:

Requirements specification, Process Discovery,

Stability Discovery, Evaluation, and Prediction, as

illustrated in Figure 1.

In the Requirements specification phase, users

define the parameters for subsequent phases. This

phase is fundamental, as it ensures that the following

steps are aligned with the user's objectives and needs.

During this phase, the user has the option to specify

process perspectives to be analyzed, select the

available process metrics to be calculated, and define

time periods for the trend analysis. For instance, a

user can select control-flow as the process perspective

to be analyzed, and choose model similarity as a

process metric, in order to perform a trend analysis of

S4BP: An Approach for Assessing Business Process Stability

211

model similarity over time. To define this trend

analysis, the user can establish monthly time periods.

Within this example, for a certain process, process

model similarity will be computed between

consecutive months, where for each month, a process

model will be considered, taking as input all the cases

which happened during that month. Furthermore, a

process engineer can choose a prediction technique

for predicting this model similarity evolution (for

instance, Simple Exponential Smoothing or linear

regression).

Figure 1. The S4BP approach.

In the second phase, titled Process discovery, the

approach foresees the discovery of process models

(for instance in the form of a Directly Follows Graph

for the control-flow perspective, or a Social Network

Analysis for the resources perspective), transforming

raw process data extracted from event logs into

actionable process models. The application of this

algorithm will facilitate the visualization of control

flow relationships, organizational roles, and

interactions among various activities, thereby

providing a deeper understanding of the current

operation of BP.

We then move on to the next Stability discovery

phase, where the selected process metrics are

computed. For instance, considering the model

similarity metric mentioned above, this phase takes

care of all associated comparisons and calculations,

as well as the computational effort to draw the

evolution of stability for this metric. In this case, the

goal is to present a trend analysis to thoroughly

examine the evolutions and adjustments made to the

models within certain time periods.

Based on these results, the user can make

informed decisions regarding which models exhibit

consistent behavior or require further attention.

Additionally, the metrics serve as a foundation for

applying an appropriate prediction technique in the

final phase, named Prediction. In this phase, the user

can estimate future process stability, allowing for

proactive adjustments and ensuring long-term

process efficiency.

5 PROTOTYPE DEVELOPMENT

AND EXPERIMENTATION

In this section, we present the prototype

development and experimentation process, which

serves to validate our approach through practical

implementation. First, we introduce the developed

mockups, which provide a visual and conceptual

representation of the proposed solution. These

mockups illustrate the key functionalities and user

interactions, offering an initial framework before full

implementation. Next, we detail the ProM plugin

implementation and experimentation, where we

describe the integration of our approach into the

ProM framework, followed by a series of

experiments to assess its effectiveness.

5.1 Developed Mockups

To validate our S4BP approach we have developed

mockups for a software application that illustrate the

appropriate user interaction. The first interface of our

prototype foresees the upload of a process event log

dataset as shown in Figure 2.

Figure 2. Import dataset interface.

The Requirements phase of S4BP approach is

prototyped in Figure 3, which involves selecting all

the parameters necessary for the subsequent phases.

Here, the interface delineates the parameters selected

by the user. The process perspective defines the

specific viewpoint to be assessed, utilizing a tailored

set of metrics. In this example, we present the

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

212

selection of the process control-flow perspective. The

selection of the associated process metrics is

contingent upon the selected perspective. For

instance, the assessment of stability from the control-

flow perspective can include (as examples) the

‘model similarity’ and 'number of traces per variant'

metrics.

Figure 3. Selecting parameters interface.

Furthermore, regarding stability trend analysis,

the user has the chance to specify the period of time

from which a process model is discovered, and

choose the desired number of process models to

assess. These discovered models will be presented in

a list format (Figure 4).

Figure 4. Process models discovery interface.

For the Stability discovery phase (Figure 5), we

illustrate it with an example of a chart, showing the

evolution of the chosen metric over time. This chart

provides a visual analysis of how the process models

have evolved over the chosen time periods.

Finally, Figure 6 presents what can be a predictive

analysis

regarding

the

chosen

process

metrics

and,

Figure 5. Stability evaluation interface.

Figure 6. Prediction stability interface.

therefore, its stability forecast. In this example, the

predicition result is shown in the form of the predicted

evolution of model similarity for the next months.

5.2 ProM Plugin Implementation and

Validation with a Running Case

In this section, we will consider the implementation

of our S4BP approach using ProM – a powerful and

widely used framework for Process Mining and

workflow analysis – and we will illustrate and

validate this implementation with a running case to

demonstrate the practical applicability of our

approach. For the latter, we used a public dataset of

the BPI Challenge initiative. This dataset has been

widely used in Process Mining research to explore

performance analysis, compliance checking, and

bottleneck identification, offering valuable insights

into real-world. It includes event logs detailing

S4BP: An Approach for Assessing Business Process Stability

213

various processes such as application submission,

document verification, offer creation, and final loan

approval or rejection and several other types. In our

work, we analyzed the “Caravan Camper” process to

assess its stability and develop predictions for the

following months. The process starts with submitting

the loan application online or in-person, specifying

the loan purpose as “Caravan/Camper”. An initial

assessment is made, including an automatic credit

check and repayment capacity evaluation, with a

manual review if inconsistencies arise. The bank then

sends loan offers, which may be adjusted by the

client, via email, postal mail, or online. Post-offer

follow-up, typically taking 15 days, and document

validation (proof of income, purchase, and insurance)

can cause delays. The final decision results in

acceptance, rejection, or cancellation, with higher

conversion rates if processed in under 30 days.

Average processing time is 22-30 days, with delays

caused by client waiting times, incomplete

documents, and offer adjustments.

For the implementation part of our approach, we

chose ProM since it is a widely recognized software

tool in the field of Process Mining, which foresees a

plug-in architecture allowing anyone to develop their

process analysis and computational programs.

For this initial implementation effort, we chose

some predetermined parameters from our S4BP

approach. For instance, for the Process discovery, we

employed the Inductive Miner algorithm, which

generates monthly models in the form of Petri nets, as

shown in Figure 7.

Figure 7. Import of 12 Petri nets models.

To evaluate the stability between the different

discovered models, we selected, as the stability

metric, the Graph Edit Distance algorithm to compute

model similarity. We then adapted the algorithm to

display the results in our prototype and to apply

evolutionary modifications. Regarding the process

perspectives, we focused on the control-flow, and for

the trend analysis, we segmented the dataset into

monthly time periods, with each segment

representing traces initiated within a specific month.

We extended the algorithm's capabilities to

support the simultaneous comparison of the multiple

discovered Petri net models, thereby overcoming the

original limitation of comparing pairs of models one

by one.

Our S4BP approach allowed us to evaluate the

stability and the fluctuations in the process and

anticipate future trends. Figure 8 shows the results of

this evaluation, comprising, as metrics, the

aforementioned model similarity, checked

sequentially between the monthly generated models,

two at a time. We can observe slight variations in

process stability, reflecting deviations over the year.

Additionally, the prototype enables users to choose

prediction techniques, such as linear regression or

ARIMA to forecast future stability for the next three

months, as shown in Figure 9. These results suggest

that the future models are expected to become

increasingly stable, though some variations are still

anticipated.

Figure 8. Comparison results, considering model similarity

over a 12-months period.

Figure 9. Stability evaluation and prediction result.

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

214

6 CONCLUSIONS

In this paper, we addressed the critical issue of

stability in BP by proposing a formal definition and a

systematic approach named S4BP that encompasses

five essential phases, each designed to allow for BP

stability assessments.

We then validated the S4BP approach through

software prototypes for user interaction and an

exploratory software implementation, using the ProM

software tool. This implementation not only

compares successive versions of the models to

evaluate their stability but also predicts their future

stability, thereby offering organizations valuable

foresight into their process dynamics.

For future work, we plan to test our prototype in a

real-case study. This step will involve evaluating and

validating our S4BP approach in a concrete

environment, incorporating predictions and real-time

monitoring. The goal is to compare the results

obtained by the prototype with those observed in the

actual situation, to confirm the reliability of our S4BP

approach. Future work will also focus on developing

a platform to automatically calculate process stability

and generate charts, particularly through dashboards.

This platform will be designed to evaluate stability

from various perspectives, not only concerning

control-flow, but also other relevant perspectives of

business processes.

ACKNOWLEDGMENTS

This work was financially supported by

Project ProM4Prod: Plataforma de Process Mining

para descoberta, medição, monitorização e

otimização de processos de produção (SI I&DT –

Empresas em Copromoção, CENTRO-01-0247-

FEDER-047242) in the scope of Portugal 2020, co-

funded by FEDER (Fundo Europeu de

Desenvolvimento Regional) under the framework of

PO CENTRO (Programa Operacional da Região

Centro).

REFERENCES

Adriano, A., Raffaele, C., Marlon, D., Marcello, L.R.,

Fabrizio, M.M., Andrea, M., Massimo, M., Allar, S.

(2018). Automated discovery of process models from

event logs: Review and benchmark. In IEEE

Transactions on Knowledge and Data Engineering,

vol. 31, no. 4, pp. 686-705.

Alejandro, B., Rebeca, C., Romero, C. (2018). Discovering

learning processes using Inductive Miner: A case study

with Learning Management Systems (LMSs).

Psicothema, 30(3), 322.

Bergaoui, N., & Ayachi Ghannouchi, S. (2024). A survey

on educational processes based on Agile, BPM, and

PM. In International Journal of Computer Information

Systems and Industrial Management Applications, 16,

Ben Haj Ayech, H., Ammar Elhadjamor, E., Ayachi

Ghannouchi, S. (2021). A systematic approach for

maintainable business process models. In Proceedings

of the 14th International Conference on Knowledge

Science, Engineering and Management (KSEM 2021).

Berti, A. (2016). Improving process mining prediction

results in processes that change over time. In Data

Analytics 2016.

Cognini, R., Corradini, F., Gnesi, S., Polini, A., Re, B.

(2016). Business Process flexibility - A systematic

literature review with a software systems perspective.

In Information Systems Frontiers a Journal of research

and innovation.

Costa, C.J. (2022). Comparing process mining tools. Ana

Catarina Ferreira Parente, ISEG – Universidade de

Lisboa, Lisbon, Portugal.

Dai, F., Xue, H., Qiang, Z., Qi, L. (2021). Refactor business

process models with redundancy elimination. In

Advances in Parallel & Distributed Processing, and

Applications.

Daoudi F, Nurcan S.(2005). A framework to evaluate

methods' capacity to design flexible business processes.

In 6th International Workshop on Business Process

Modeling. vol. 12; p. 1-8.

Gezani, R.M., Seeletse, S.M. (2015). Quantifying Business

Process Optimization using Regression. In American

Journal of Applied Sciences, 12(12), 945-951.

Hompes, B.F.A., Buijs, J.C.A.M., van der Aalst, W.M.P.,

Dixit, P.M., Buurman, J. (2015). Discovering deviating

cases and process variants using trace clustering. In

Proceedings of the 27th Benelux Conference on

Artificial Intelligence (BNAIC).

Huang, Z., Kumar, A. (2012). A study of quality and

accuracy tradeoffs in process mining. In INFORMS

Journal on Computing, 24(2), 311–327.

Jensen, W., Szarka, J. (2018). Stability assessment with the

stability index. In Quality Engineering.

Kelly, D. (2006). A study of design characteristics in

evolving software using stability as a criterion. In IEEE

Transactions on Software Engineering, 32(5).

Kim, J., Comuzzi, M. (2021). Stability metrics for

enhancing the evaluation of outcome-based business

process predictive monitoring. In IEEE Access.

Kherbouche, M.O. (2013). Contributions à la gestion de

l’évolution des processus métiers. In Thèse en

informatique, Université du Littoral Côte d’Opale.

Pandey, S., Nepal, S., Chen, S. (2012). A test-bed for the

evaluation of business process prediction techniques. In

7th International Conference on Collaborative

Computing: Networking, Applications and

S4BP: An Approach for Assessing Business Process Stability

215

Worksharing (CollaborateCom). IEEE. Orlando, FL,

USA.

Lavanya, M.U., Talluri, M.S.K. (2015). Dealing with

concept drifts in process mining using event logs. In

International Journal of Engineering and Computer

Science, 4(7), 13433-13437.

Lin, H., Qu, T., Li, L., Tian, Y. (2020). The paradox of

stability and change: A case study. In Chinese

Management Studies: Volume 14 Issue 1.

Luengo, D., Sepúlveda, M. (2012). Applying clustering in

process mining to find different versions of a business

process that changes over time. In Lecture Notes in

Business Information Processing, 153-158.

Mejri, A., Ayachi-Ghannouchi, S., & Martinho, R. (2018).

A quantitative approach for measuring the degree of

flexibility of business process models. Business

Process Management Journal, 24(4), 1023-1049.

Mejri, A., Ayachi-Ghannouchi, S., & Martinho, R. (2024).

Flexibility-based selection of paradigms and tools for

BP modelling and execution. International Journal of

Business Process Integration and Management, 11(3),

165-178.

Pesic M, Van der Aalst WM. (2006).A declarative approach

for flexible business processes management. In:

Business Process Management Workshops. Springer;

p. 169-180.

Ramirez, S. M., Cortes, K., Ocharan-Hernandez, J. O., &

Sanchez Garcia, A. J. (2018). Software stability: A

systematic literature review. In Proceedings of the 6th

International Conference in Software Engineering

Research and Innovation. School of Statistics and

Informatics, Universidad Veracruzana, Xalapa,

Veracruz, Mexico.

Regev G, Soffer P, Schmidt R. (2006). Taxonomy of

Flexibility in Business Processes. In: BPMDS; 2006.

Pesic M, Van der Aalst WM. (2006). A declarative

approach for flexible business processes management.

In Business Process Management Workshops.

Springer; p. 169-180.

Schonenberg H, Mans R, Russell N, Mulyar N, van der

Aalst WM. (2008). Towards a Taxonomy of Process

Flexibility. In CAiSE forum. vol. 344; p. 81-84.

Saint, J., Fan, Y., Sing, S. (2021). Using process mining to

analyse self-regulated learning: A systematic analysis

of four algorithms. In LAK21, April 12–16, 2021,

Irvine, CA, USA.

Schonenberg H, Mans R, Russell N, Mulyar N, van der

Aalst WM (2008). Towards a Taxonomy of Process

Flexibility. In CAiSE forum. vol. 344;. p. 81-84.

Van der Aalst, W., Buijs, J., van Dongen, B. (2011a).

Improving the representational bias of process mining

using genetic tree mining. In SIMPDA Proceedings

2011.

Van der Aalst, W.M.P., Adriansyah, A., De Medeiros,

A.K.A., Arcieri, F., Baier, T., et al. (2011b). Process

mining manifesto. In Proceedings of the 9th

International Conference on Business Process

Management, 169-194.

Yau, S.S., Collofello, J.S. (1980). Some Stability Measures

for Software Maintenance. In IEEE Transactions on

Software Engineering, 6(6), 545-552.

Yau, S.S., Collofello, J.S. (1985). Design Stability

Measures for Software Maintenance. In IEEE

Transactions on Software Engineering, 11(9), 849-856.

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

216