Towards a Conceptual Model for Undesired Situation Detection through

Process Mining

Matheus Friedhein Flores

, Den

ılson Ebling

, Jonas Bulegon Gassen

, Vin

ıcius Maran

and Alencar Machado

Col

egio Polit

ecnico, Universidade Federal de Santa Maria, Santa Maria, Brazil

Antonio Meneguetti Faculdade, Santa Maria, Brazil

Laboratory of Ubiquitous, Mobile and Applied Computing, Universidade Federal de Santa Maria, Cachoeira do Sul, Brazil

Keywords:

Process Mining, Situation Detection, Control Charts, Proactive Actions.

Abstract:

As technology advances, recent research propose solutions to monitor and control organizational processes,

aiming to maximize efﬁciency and productivity and minimize the loss of resources involved in the execution

of processes, whether human or technological, in addition to maintaining a controlled environment so that the

objectives of the organizations are achieved, that is, the satisfaction of their customers. For this, historical

information contained in the event log is frequently used, related to the execution of processes in the orga-

nizational environment. These information serves as a basis for controlling the environment, preventing the

occurrence of unwanted situations. In this context, this paper presents a model for detecting situations of in-

terest in the organizational environment through event logs, making it possible to initiate proactive actions in

the face of these situations, resulting in a Web application provided by interfaces that validate the purpose of

the article. Beyond the scenario, an event log related to the execution of a real process was tested. By means

of control charts, it is possible to view (using time parameters) the delay in the execution of the process, which

may be related to a situation of interest.

1 INTRODUCTION

In contemporary organizations there might be consid-

erable wastes of ﬁnancial and human resources. The

lack of planning and monitoring of existing processes

may have a great impact on them. These factors may

inﬂuence products or services towards low quality

and loss of customers, among other factors that im-

pact the direct form of the organization. To address

these problems, organizations are increasingly invest-

ing in developing, expanding and improving internal

processes through the use of areas such as: log min-

ing, process control and monitoring (Van Der Aalst,

2011) (Saylam and Sahingoz, 2013). Managing pro-

cesses improves organizations efﬁciency and allow

them to reach their objectives, providing agile inter-

actions with its customers. In addition to improving

the quality and development of the work carried out, it

minimizes internal problems and waste of resources,

whether human or ﬁnancial. Usually the manage-

ment of processes is performed by means of software,

which helps with versioning and execution (Laudon

and Laudon, 2015).

Based on these processes, it is possible to iden-

tify desired behavioral parameters. For example, exe-

cution time of activities in a process can be used to

identify if the process is running within the estab-

lished quality limits of the organization in the pro-

vision of any service. In order to manipulate the in-

formation provided by the software that executes the

processes, process mining research area arises, which

proposes to suggest procedural improvements, dis-

cover control ﬂow, performance analysis, compliance

and monitoring tasks, characteristics that impact the

internal and external environment of an organization

(Van Der Aalst, 2011).

According to Van Der Aalst (2011), process min-

ing is a research area that focuses on the extraction of

information about the behavior of business processes

executions. It uses log of events recorded by informa-

tion systems available in organizations. Event logs,

while not adding much value when viewed individ-

ually, when analyzed together, can be useful to un-

derstand or even improve and optimize the processes

that make up a business environment. By analyzing

the logs in sets, it is possible to deﬁne execution pat-

Flores, M., Ebling, D., Gassen, J., Maran, V. and Machado, A.

Towards a Conceptual Model for Undesired Situation Detection through Process Mining.

DOI: 10.5220/0009564408090816

In Proceedings of the 22nd International Conference on Enterprise Information Systems (ICEIS 2020) - Volume 2, pages 809-816

ISBN: 978-989-758-423-7

809

terns and consequently situations of interest in the en-

vironment in which the user is involved. The model

proposed by Machado (2017) aims to manage an en-

vironment in which situations occur. Such situations

generate events and the detection of patterns related to

these events characterizes situations of interest, which

can be ignored or not. In case of an unwanted situa-

tion, the model performs actions to control the envi-

ronment and to avoid damage related to the monitored

context.

It is worth mentioning that many organizations

ﬁnd difﬁculties in monitoring and controlling their

processes at run time, by the fact that they produced

information (logs), based on the execution of their

processes, and do not know such as to best use this in-

formation. Inﬂuencing the ﬁnal result of the product

or service provided. This work presents a model to

monitor and control the organizational environment,

in order to avoid situations that negatively impact

business processes. The proposed model is an exten-

sion of the model of proactive actions proposed by

Machado (2017), with characteristics of the environ-

ment being added to the model so that it can be con-

trolled, minimizing the occurrence of undesired situa-

tions. Taking into account the times and performance

in the execution of process in the business environ-

ment, being able to act proactively or reactive in the

face of environmental situations, if necessary.

The article is structured as follows: In Section 2,

we present the main concepts found in the literature.

The model developed in this work is presented in sec-

tion 3. Section 4 presents a case study using the model

and is followed by section 5 where results and discus-

sions are presented. Finally, in section 6, we draw

some conclusions and indicate points for future work.

2 BACKGROUND AND RELATED

WORK

This section describes concepts that serve as a the-

oretical basis for the development of the work, such

as: Process Mining, Log, Control Charts and Model

to Proactive Identiﬁcation of Situations of Interest,

besides Related Work. According to Van Der Aalst

(2011), process mining aims to extract knowledge

from data generated by the execution of processes in

information systems used by organizations. It seeks

to discover, monitor and improve existing real pro-

cesses. Process mining emerged in the 1990s, with

the ﬁrst works focusing on mining process models

in software engineering event logs (Cook and Wolf,

1995).

Process mining is an area of research positioned

between the areas of artiﬁcial intelligence, data min-

ing, process modeling and analysis; therefore, the

growing interest in the area can be justiﬁed by the

fact that more and more events are recorded. (Van

Der Aalst, 2016) (Burattin, 2013). The increasing

volume of digital information related to processes in

organizations allows the registration and analysis of

their events. Any step or operation of a process or sys-

tem can be seen as an event. (Van Der Aalst, 2012a).

In other words, process mining is important and

efﬁcient, as it is capable of converting historical infor-

mation (logs), related to a process, enabling the orga-

nizations specialists to view, monitor and control what

is really happening in the execution of processes. An

event log consists of the record of events that occurred

during the functioning of an organization, and the

storage of this record is carried out mainly by infor-

mation systems (Van Der Aalst, 2011)(Glavan, 2011).

Events are considered as a tuple containing the fol-

lowing ﬁelds (Van Der Aalst, 2011): (i) ID: identiﬁ-

cation; (ii) Timestamp: activity start date and time;

(iii) Activity: Activity description; (iv) Resource: re-

sponsible for the activity; (v) Cost: activity cost; (vi)

Extra data: additional information about an activity.

The log data can be located in a distributed and in-

complete way, inconsistent with the reality and con-

taining outliers (noise). Additionally to quiet activ-

ities that are performed in the company and depend

on third parties, therefore not being registered in the

system (Van Der Aalst, 2012b).

In order to measure the parameters provided by

the logs, control charts can be used. Therefore, it is

possible to deﬁne upper and lower limits, so that met-

rics and can identify the occurrence of anomalies in

the stored records. In addition to monitoring and con-

trolling the actions involving these records. Accord-

ing to Oliveira (2013), control charts are tools used to

monitor the performance of a process, based on char-

acteristics that they call control limits. These limits

are known as (i) upper line (upper control limit - LSC)

and (ii) lower line (lower control limit - LIC), in ad-

dition to (iii) central line (central limit - LC).

According to Oliveira (2013), when all the sam-

ple points are within the limits of control, it is con-

sidered that the process is ”under control”. How-

ever, if one (or more) points are positioned outside

the imposed control limits, there is evidence that the

process is ”out of control” and that an investigation

into the occurrences and corrective actions are needed

to detect and eliminate special causes in the process.

Therefore, after sample analysis, it is possible to de-

ﬁne whether a situation in the environment is of inter-

est, that is, characterizes an uncontrolled environment

(Machado, 2017).

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

810

The work presented in Machado (2017) deals with

the identiﬁcation and management of situations of in-

terest, whether unwanted or not, in intelligent envi-

ronments. It seeks to perform actions (using web ser-

vices) in the face of such situations. The work pro-

poses an approach for systems to act in these situa-

tions, trying to eliminate or soften their impact. In ad-

dition, intends to expand their ability to manage new

unwanted situations that arise in the user’s life envi-

ronment over time.

Machado (2017) deﬁnes that situations are com-

posed of events, being monitored by the middleware

that provides some type of service. These events take

place in a valid time window called a sliding win-

dow, and the correlations of these events can high-

light a situation. If a situation of interest occurs and

it is unwanted, actions can be taken to control the en-

vironment, which can be reactive or proactive. For

more information about the model and concepts, see

(Machado, 2015).

An event has a name, being characterized by a

type (internal or external) and has a time within

∆

(ﬂow of events monitored as data time), and a set of

semantic relations of the context R. When an event is

not produced by a simple entity (raw data from a sen-

sor or simple record of the consumption of a service),

it must also contain a standard (p) for its detection

(Machado, 2017) (1):

Event : (name,type,time, {R}, p) (1)

As Machado (2017) describes, events evaluation

(a standard) can make the system identify the likeli-

hood of an unwanted situation to happen in the fu-

ture. Other works, such as Forkan (2015) and Coro-

nato (2014), deﬁne that in order to understand fu-

ture changes in user behavior, a system must consider

proactive actions and current events. So, an approach

can understand the user’s behavioral deviation using

pattern recognition models. Seeking to validate the

objective of this work, as well as to compare it with

the literature, arise the work proposed by Mannhardt

and Landmark (2019). It aims to apply process min-

ing techniques to the rail trafﬁc control event logs

(case study), allowing to investigate the quality of de-

cisions made by those responsible for the railroad.

The work proposed by Tax and van der Aalst

(2019), conceptualizes techniques to ﬁlter chaotic ac-

tivities from event logs. The case study used by

the authors constitutes a collection of seventeen (17)

event logs derived from the execution of real life busi-

ness processes. Based on the authors’ results, they ar-

gue that chaotic activities affect the quality of process

models.

Polato (2018) proposes a model to predict times

and deal with unexpected behaviors based on the ad-

ditional data present in the event logs. For this, the

authors use machine learning techniques to train al-

gorithms and detect unexpected behavior.

Based on the works mentioned above, it is pos-

sible to analyze that all works deal with some fact

that occurred as a chaotic situation, similar to the

unwanted situation, which may be infrequent or fre-

quent, depending on the case. However, none of the

aforementioned works contains characteristics that

make it possible to take actions in relation to situa-

tions of interest, minimizing their impact on the re-

lated environment. The present work describes, in the

next section, a model that allows controlling and mon-

itoring the processes, detecting situations of interest

and, if undesired, executing deﬁned reactive actions.

3 A SITUATION DETECTION

MODEL FOR PROCESS

MONITORING AND CONTROL

In the work of Machado (2017), the focus of the re-

search refers to assisted living environments, manag-

ing situations that occur in that environments, with

the help of several sensors to capture contextual infor-

mation of users with cognitive problems or with ad-

vanced age. But, as conceptualized by Machado, the

model was proposed for generic environments, that is,

it can be used in any context.

In the organizational context, the model from

Machado (2017) can be extended for the control and

monitoring of processes. Considering how to perform

actions on behalf of the organization, based on infor-

mation generated during the execution of processes,

in addition to being able to detect and interpret sit-

uations that occur in the environment. Based on the

event log (Section 2 - log), which contains the his-

torical information stored about actions that have oc-

curred, it is possible to deﬁne parameters indicating

when a process is exposed to some situation of inter-

est. These parameters are related to the execution of

the processes. In this work, the time and noise param-

eters are taken into account. To improve understand-

ing of the business and the applied model, Figure 1

illustrates an overview of the business and the context

around all actions in the process.

• a) Figure 1 illustrates the environment in which

the processes are executed, starting with the

client’s negotiation with the business manager

(team responsible for the organization’s negotia-

tions). In this negotiation, some information is

generated, in the presence of the customer and

without the presence of the customer, such as ﬁll-

Towards a Conceptual Model for Undesired Situation Detection through Process Mining

811

Figure 1: Overview of the organizational context.

ing out records or updating personal information.

• b) The process starts in this negotiation, but it is

not limited to this interaction between the client

and the business manager, but passes to the inter-

nal environment of the organization, being pro-

cessed between different activity until it is re-

jected or accepted. In each activity where the pro-

cess is processed, various information are gener-

ated, for example, who was responsible for pro-

cessing another activity, what time it occurred,

such as long it took to process, whether the pro-

cess was idle or not. All of this information gen-

erated are events and each event is saved in the

database, resulting in the event log of the process.

• c) These processes are controlled based on the

event log, taking into account the time factor of

the processes execution.

• d) Using the control chart(Section 2), it is possible

to deﬁne upper and lower bounds to compare pro-

cesses with execution times within the expected

time and special causes, times that exceed the im-

posed limits and allow to indicate whether a pro-

cess is in progress . a controlled or uncontrolled

environment due to the detection of event patterns.

• e) Based on the work of Machado (2017), which

consists of managing smart environments using

middleware and applications, fed by information

generated by various sensors embedded in the en-

vironment, aiming to control and monitor to min-

imize or prevent the user involved from putting

himself in unwanted situations, however, in this

article, the information is provided by the event

logs and managed by the middleware provided by

an application.

The objective of the present work is similar to that

proposed by Machado (2017), however, here the in-

formation is provided by log events and not by sen-

sors. The conceptual model proposed by Machado

can be applied and extended to the context of this

paper. For this, some concepts were added to meet

the demands and needs of the organizational environ-

ment, illustrated in the ﬁgure 2, such as:

• (a) Legacy systems, related to third-party sys-

tems used for customer service and business pro-

cess organization. Software to automate routines

and bureaucratic procedures, with computerized

services that provide the data to be manipulated

and managed, aiming to control the organizational

environment.

• (b) Noise, incorrect information that can nega-

tively inﬂuence time calculations, that is, for un-

foreseen reasons, the information can be mislead-

ing, for example, the person responsible for pro-

cessing the process in a certain sector, went on va-

cation, but the process was not passed on to any-

one, becoming idle, consequently increasing the

execution time and negatively impacting the orga-

nization’s image in relation to the client.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

812

Figure 2: Proactive model for extended business processes.

As described in concept (b), data provided by

third-party software saved in logs may present errors,

due to logic, commercial rule, technology or other un-

foreseen factors. To minimize errors when using in-

formation from the logs, all data must be ﬁltered and

analyzed, so that they can be processed as execution

occurs.

For each data in the log, when it is ﬁltered, it is

subjected to a base calculation to be categorized as

noise or not. The base calculation is equivalent to an

accepted time value above the average time deﬁned

by the historical data of the process under analysis. If

the average time is 10 (ten) minutes for the execution

of the process, only events with a time greater than the

average time in 3 (three) times the average will be cat-

egorized as noise (2), that is, if process Y is related to

a situation of interest X, which takes an average of 10

(ten) minutes to execute, information over time of 30

minutes is disregarded and is not taken into account

in detecting the situation of interest.

Noise = time > (3 ∗ average) (2)

The information, categorized as noise, is ignored

because, if taken into account, they end up interfering

in the ﬁnal result of the moments when the process is

executed, causing errors and inconsistent information

in the sample being controlled and monitored. In or-

der to differentiate between with and without noise,

the noise attribute was added to the event proposed

by Machado (2017) (3), so that it is possible to dis-

tinguish events that can be analyzed from those that

should be discarded.

Event : (name,type,time, noise, {R}, p) (3)

Given the logs, having the middleware and a sys-

tem to monitor and control running processes, it is

possible to use the time parameter, deﬁned by means

of historical information, to audit the processes and

discover situations of interest. In addition to deter-

mine actions for such situations. For this, the sys-

tem provides an algorithm with equations and prob-

abilities that manipulate the times of execution of a

process. Such as calculation of times by process, in

which it interprets individual values to deﬁne param-

eters to be used in comparisons of execution times,

saved in the event logs. During execution, the pro-

cess is monitored and controlled by the diffused sys-

tem that ﬁlters data equivalent to noise and detects

patterns in times that exceed the limits deﬁned by the

control chart for a certain period, possibly character-

izing an unwanted situation. To validate the objective

and the extended model, the next section describes the

case study with an empirical scenario, which charac-

terizes an organizational environment.

4 CASE STUDY SCENARIO

In order to validate the model described in this work,

a scenario that reﬂects the routine events of a banking

organization was elaborated. The scenario consists of

a business environment in which data is created, tak-

ing into account the interactions of a customer, who

requests a product, and an organization, which sells

that product. Let us call the situation produced in

this scenario as ”Customer dissatisfaction” initiated

by the event ”credit denied”.

The following scenario was considered to apply

the extension of Machado (2017) to another context.

John, 26, is a development analyst and works for a

Towards a Conceptual Model for Undesired Situation Detection through Process Mining

813

private technology company. One day, John decided

to buy a vehicle, but he did not have the total amount

necessary for the purchase. Given this fact, he chose

to go to a bank and apply for a credit to complete the

required amount.

The organization where John is a customer is one

of the best known in the banking sector. It has almost

all of its processes computerized and well-structured,

with the aim of always meet customer needs. Previ-

ously to the decision of buying a vehicle, John had

already negotiated with the organization. In the prior

negotiation, he asked for a construction credit (res-

idential credit line) to ﬁnish his apartment and had

this credit approved. In such case, the ”credit denied”

event was not detected, consequently, not causing the

situation ”customer dissatisfaction”.

Satisﬁed, John returns to the organization in order

to request credit for the purchase of the desired vehi-

cle, in a hurry to close the deal with the seller. Histor-

ically, the vehicles credit line moves faster than other

credit lines in this organization. However, in John’s

request, there were some ﬂaws in the execution of the

process. Due to the delay in the execution of the pro-

cess, John decided to change the vehicles credit line

to the simple credit line, used when users have been

customers for some time. But it was different from

the ﬁrst contact with the organization and the usual

cases of the same process. When John requested the

related credit with the simple credit line, his request

was denied, causing the organization to experience an

unwanted ”customer dissatisfaction” situation.

In order for the unwanted situation to be detected,

the ”credit denied” event initiates the unwanted situ-

ation of ”customer dissatisfaction”; however, this sit-

uation occurs only when another related event is de-

tected, when Peter (commercial manager) offers an-

other line of credit as an option for John to leave

with the approved credit, being ”offer another line of

credit” the detected event.

The occurrence of these events negatively affects

the organization, as it is the opposite of their strategic

objective: customer satisfaction. In view of this situ-

ation, the organization has deﬁned some measures to

manage its processes, seeking to avoid situations like

these. One measure was to deploy a middleware and

a comprehensive application to monitor and manage

the business environment. All information resulting

from the customer interactions with the organization

is managed by the middleware and the application,

as the two together have the ability to assist the or-

ganization in the execution of its processes. It helps

to control the emergence of situations, as well as to

deﬁne proactive and reactive actions when such situ-

ations occur.

In the case of John, when a situation of inter-

est ”credit denied” is detected, the application im-

mediately asks the middleware to send a notiﬁcation

through some communication channel to the organi-

zation’s manager, taking into account that he takes

some action in relation to the event that occurred, in

order to minimize the damage for both sides.

The scenario described above demonstrates that a

situation of interest was detected, given this fact, ex-

tending the Machado (2017) model, a application trig-

gers reactions to control the environment in question,

as described in the next section.

4.1 Behavior for Credit Denied

Situations

When installed, the comprehensive application in-

forms the middleware based on the event logs, their

context and their situation of interest, according to

the parameters deﬁned for the comparisons made in

the extended model (section 3), disregarding noisy

data. If a running process detects that John received

the denied credit (the ea1 event was detected), the

”customer dissatisfaction” situation is initiated by the

application and the application chooses an action to

be triggered, with action a1 sending a notiﬁcation

through a communication channel to activate the re-

sponsible personnel.

Current Situa-

tion:

< (customer dissatisfaction),

(ea1), (ea2) >

Event name ea1

Description credit denied

Type Internal

Noise false

{R}

<John, GoToAgency,

RequestCredit> ∧

<RequestCredit,

CheckDocuments,

ValidDocumentation>∧

< RequestCredit, = ,

InValidDocumentation >

⇒ <John, creditRequested,

denied>

Pattern

SELECT e FROM

PATTERN [every e=Event

(name= ’ea1’)]

With the continuous monitoring of the environ-

ment in relation to the events that occur, after send-

ing the notiﬁcation, the system understands that the

situation of ”customer dissatisfaction” has ended.

As shown above, if the ea2 event is detected, John

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

814

Event name ea2

Description credit denied

Type Internal

Noise false

{R}

<John ,GoToAgency,

RequestCredit>

∧ <RequestCredit,

CheckDocuments,

ValidDocumentation>

∧<ValidDocumentation,

containsMinimumRequirements,

RequirementsOk>

∧<RequestCredit, =

, creditReleased>

=<John, RequestCredit,

released>

Pattern

SELECT e FROM PATTERN

[every]e=Event (name= ’ea2’)]

has his credit released, there is no situation of ”cus-

tomer dissatisfaction”. According to the model pro-

posed by Machado (2017) (section 2), the actions are

activated corresponding to the events generated, as

described in the scenario, action a1 generates event

a1, action a2 generates event a2 and so on. The ac-

tions for the ”customer dissatisfaction” situation are

described below, being declared as Action (name, Sp,

Se, E) (Machado, 2015). Each action generates an

event that the application runs to control the envi-

ronment and the detected situation, in order to avoid

losses.

Name: Notiﬁcation by whatsapp after ea1;

{Sp} credit denied;

Services used {Se}

Se notifyWpp

{I}

Device,

“checkDocumentation”;

{O}:

DeviceMobile,

SendMessageWhatsApp

Services resulting in action {E}

Name a1

Type Internal

Noise false

{R}

notifyApp

ActionAutomated,

haveService,

notifyService>

∧ notiﬁcationService,

aTypeOf,

interaction User>

Pattern

SELECT e FROM PATTERN

[every e=Event(name =’a1’)->

and Event(name =’ea1’))]

5 RESULTS AND DISCUSSION

Based on the scenario described above, where the en-

vironment in which the processes are executed is con-

ceptualized, a prototype was developed, which is in

the testing phase. The web application intends to as-

sist the monitoring and control of processes. It was

implemented in an organization, being tested in the

real environment with customers. The data is pro-

vided by internal software used by the organization to

manage its daily tasks, which are stored in the event

log from where the application consumes informa-

tion.

The application consists of some interfaces, one

of which is illustrated in Figure 3, related to the ex-

ecution schedule of a single process from ”General

Credit” type, with the times and their respective dates

as the stored execution occurred in the event log. The

application has an interface to perform searches and

generate reports related to the execution times of the

processes based on the event logs. As shown in ﬁgure

3, there are some ﬁelds that allow you to specify the

period of the start and end dates, the type of process,

in addition to the area and the activity in which it is

found, information that helps to seek more accurate

data about the process.

Figure 3: Graphical interface for monitoring and controlling

process.

When entering the chosen ﬁelds for the search, a

graph related to the control charts is displayed, show-

ing the estimated times on the X axis and the date

on which the fact occurred in Y, as well as lines and

points that characterize the average time (blue dotted

line), the time peaks of some processes that occurred

longer than expected (red line) and the expected time

(black line).

It is possible to observe that most of the points

are above the expected schedule and another one even

higher than the average, being very extreme cases. In

cases of very high times, the analysis of special cases

Towards a Conceptual Model for Undesired Situation Detection through Process Mining

815

would be made, being calculated with the accepted

percentage, that is, in relation to ﬁgure 3, cases (a)

and (b) would be categorized as noise, greater than

3 times the average of all times in the sample, being

removed from the analysis to detect the situation of

interest. However, it is worth mentioning that, with

the period of execution of the inserted process, with

start and end date, it turns out that the times occurred

exceed the expected time per process, detecting the

undesired situation resulting from the delay in the ex-

ecution of the process, initiating proactive actions by

the application, which can be, as described in the case

study, a warning message to the person responsible

for the process.

6 CONCLUSION

In this work, a model that extends to that proposed by

Machado (2017) was presented. Using the extended

model, it was possible to monitor and control the pro-

cesses, to minimize losses due to unwanted situations.

The contributions of the work can be described

as: (i) it is possible to detect an intersection situation

based on historical data provided by the software that

assist in the execution of organizational processes; (ii)

reactive actions in the face of unwanted situations that

meet the objectives imposed by the organization and

control the organizational environment; (iii) improve-

ment in the ﬂow of process executions, reducing the

number of occurrences and unwanted situations, in-

creasing efﬁciency in the business environment.

Thus, we understand that the expanded model,

together with the areas of mining, monitoring and

process control, have much to add to organizations

concerned with developing and improving their pro-

cesses. Aiming to gain efﬁciency, productivity and

safety, seeking to provide services or products with

added quality to their customers.

REFERENCES

Burattin, A. (2013). Applicability of process mining tech-

niques in business environments. Doctoral diss., alma.

Cook and Wolf, A. (1995). Automating process discovery

through event-data analysis. ICSE ’95, pages 7382,

New York, NY, USA, 1995. ACM. 22, 27, 34.

Coronato, A.; Paragliola, G. (2014). Towards a Personal

Health Records system for patients with Autism Spec-

trum Disorders. In: Computational Intelligence in

Healthcare and e-Health. Proceedings Symposium

Series on Computational Intelligence.

Forkan, A. R. M. e. a. (2015). A context-aware approach

for long-term behavioural change detection and ab-

normality prediction in ambient assisted living. Trans-

actions on Management Information Systems (TMIS),

3(2), 1-17.

Glavan, L. (2011). Understanding process performance

measurement systems. Business Systems Research,

2(2), 25-38. 30, 35, 39.

Laudon, K. C. and Laudon, J. P. (2015). Management In-

formation Systems: Managing the Digital Firm Plus

MyMISLab with Pearson eText. Prentice Hall Press,

Upper Saddle River, NJ, USA.

Machado, A. (2015). Sensibilidade

a situac¸

ao em ambi-

entes de viv

encia assistida: uma abordagem reativa,

proativa e extens

ıvel.

Machado, Alencar, e. a. (2017). Reactive, proactive, and

extensible situation-awareness in ambient assisted liv-

ing. Systems with Applications 76 (2017): 21-35.

Mannhardt, F. and Landmark, A. D. (2019). Mining railway

trafﬁc control logs. Transportation research procedia

37.

Oliveira, C. C., e. a. (2013). Manual para elaborac¸

ao de

cartas de controle para monitoramento de processos

de medic¸

ao quantitativos em laborat

orios de ensaio.

Instituto Adolfo Lutz, S

ao Paulo.

Polato, Mirko, e. a. (2018). Time and activity sequence

prediction of business process instances. Computing

100.9.

Saylam, R. and Sahingoz, O. K. (2013). Process mining

in business process management: Concepts and chal-

lenges. In International Conference on Electronics,

Computer and Computation (ICECCO).

Tax, Niek, N. S. and van der Aalst, W. M. (2019). Discov-

ering more precise process models from event logs by

ﬁltering out chaotic activities. Journal of Intelligent

Information Systems.

Van Der Aalst, W., A. A. D. M. A. K. A. A. F. B. T. B. T.

. . B. A. (2012a). Process mining manifesto. In 9th

International Conference on Business Process Man-

agement, BPM 2011P (pp. 169-194). Springer Verlag.

12, 13, 14, 15, 16, 17,18, 19, 20.

Van Der Aalst, W. (2011). Process mining: discovery,

conformance and enhancement of business processes

Vol(2). Heidelberg: Springer.

Van Der Aalst, W. (2012b). Process mining: Overview and

opportunities. ACM Transactions on Management In-

formation Systems (TMIS), 3(2), 1-17.

Van Der Aalst, W. (2016). Data science in action. In Pro-

cess mining. Springer, Berlin, Heidelberg.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

816