Emotional States Management for an Advanced Intelligent Tutoring

System

Domenico Redavid

1 a

, Stefano Ferilli

2 b

, Liza Loop

and Liudmyla Matviichuk

4 c

Economics and Finance Department, University of Bari, Largo Abbazia S. Scolastica, Bari, 70124, Italy

Computer Science Department, University of Bari, Via E. Orabona 4, Bari, 70125, Italy

LO*OP Center, Inc., 16511 Watson Rd, Guerneville, 95446, CA, U.S.A.

Department of Informatics and Computing Tools, T. H. Shevchenko National University ”Chernihiv Colehiu”,

53, Hetmana Polubotka Str., Chernihiv, Ukraine

Keywords:

Intelligent Tutoring System Framework, Ontologies, Emotional States.

Abstract:

One of the result of the application of Artiﬁcial Intelligence (AI) to e-learning environments are Intelligent

Tutoring Systems (ITSs). A crucial aspect in the ﬁeld of e-learning concerns emotional states, which impor-

tance is increasingly felt also at the level of educational institutions. On the other side, the Management of

moods at Information Technology level is becoming more and more important because enable new scenarios

where new innovative applications can be proposed. In this paper is described a possible framework able to

manage emotional states that could be adopted as a solution to address the Personal, social and learning to

learn competence, one of the eight key competences lifelong learning indicated by EU COUNCIL.

1 INTRODUCTION

Intelligent Tutoring Systems, or ITSs, are a conse-

quence of the application of Artiﬁcial Intelligence

(AI) to e-learning environments. The ITS vision can

be summarised as the creation of a personalised tutor

for each student able to provide them with exactly the

support they need to get the most out of their learning

experience. Unlike a dedicated human tutor, the ITS

would also be able to take into account many more

parameters and data and thus be fully compliant with

the goals and objectives (strict or otherwise) of each

individual supported user. Being able to also report

to the learner on the effectiveness and efﬁciency of

the learning activities would increase his or her ca-

pability to acquire skills and/or knowledge. Students

would also be empowered to share this information

with their teachers, thereby enhancing teacher effec-

tiveness. Recently, KEPLAIR (Knowledge-based En-

vironment for Personalised Learning using an Artiﬁ-

cial Intelligence Recommender) (Ferilli et al., 2021;

Ferilli et al., 2022) has been proposed. It is an ITS

designed to make pervasive use of artiﬁcial intelli-

https://orcid.org/0000-0003-2196-7598

https://orcid.org/0000-0003-1118-0601

https://orcid.org/0000-0002-2046-6153

gence to perform its tasks. Although KEPLAIR was

designed for independent learners, due to its advanced

features it can also be used in contexts such as com-

pulsory schooling to involve students in the choice of

study materials and, as in the case proposed in this pa-

per, in the choice of learning paths. Indeed, for some

years, there has been an in-depth discussion at the Eu-

ropean level on the subject of the skills that individ-

uals must acquire in order to ensure their full devel-

opment. This is a pivotal issue, with cascading im-

plications that affect the issues of training, education

and orientation towards work and social welfare. The

result of this process was the elaboration of the eight

European key competences, which the Member States

of the European Union are called upon to transpose,

facilitating their acquisition by all citizens. The refer-

ence text is the Recommendation on key competences

for lifelong learning

approved by the European Par-

liament on 22 May 2018. The Reference Framework

sets out eight key competences:

• Literacy competence,

• Multilingual competence,

UE, COUNCIL RECOMMENDATION of

22 May 2018 on key competences for life-

long learning - https://eur-lex.europa.eu/legal-

content/EN/TXT/HTML/?uri=CELEX:32018H0604(01)

Redavid, D., Ferilli, S., Loop, L. and Matviichuk, L.

Emotional States Management for an Advanced Intelligent Tutoring System.

DOI: 10.5220/0011590200003335

In Proceedings of the 14th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2022) - Volume 2: KEOD, pages 253-259

ISBN: 978-989-758-614-9; ISSN: 2184-3228

253

• Mathematical competence and competence in sci-

ence, technology and engineering,

• Digital competence,

• Personal, social and learning to learn competence,

• Citizenship competence,

• Entrepreneurship competence,

• Cultural awareness and expression competence.

In particular, Personal, social and learning to learn

competence is related to the ability to self-reﬂect,

to manage time and information effectively, to work

with others constructively, to remain resilient and to

manage one’s own learning and career. It includes

the ability to deal with uncertainty and complexity, to

learn to learn, to support one’s physical and emotional

well-being, to maintain physical and mental health,

and to be able to lead a health-conscious and future-

oriented life, to empathize and to manage conﬂict in

an inclusive and supportive context. Emotional well-

being is closely interconnected with the recognition

of emotional states.

The objective of this position paper is to outline

a possible IT framework able to manage emotional

states that could be exploited as a solution for a prac-

tical case of implementation of the EU recommenda-

tions on key competences, the Italian case on PCTO

2 BACKGROUND

In this section we report the related work and some

speciﬁc prototypes implementation that will be used

to implement the proposed ITS framework.

2.1 Related Works

The recognition of emotional states (also known as

moods) falls within the scope of affective computing,

an interdisciplinary research ﬁeld spanning computer

science, psychology, and cognitive science. Mood

has a lower intensity but a longer duration than emo-

tion since it is an emotional state that affects the ex-

perience and behavior of a person. To the best of

our knowledge, unlike standard emotions where vari-

ous approaches have been proposed (Imani and Mon-

tazer, 2019), (Yadegaridehkordi et al., 2019), frame-

work proposal for the management of moods in ed-

ucation from IT perspective has not yet been inves-

Linee guida dei percorsi per le competenze

trasversali e per l’orientamento (PCTO), MIUR,

4/9/2019 - https://www.miur.gov.it/web/guest/-/linee-

guida-dei-percorsi-per-le-competenze-trasversali-e-per-l-

orientamento

tigated in depth. In (Abaalkhail et al., 2018) is pre-

sented a survey on ontologies for the representation

of affective states and their inﬂuences including pro-

posal for moods ontologies. Indeed, the paper denotes

that there are only few ontologies proposal that target

mood and human inﬂuence. In (Bolock et al., 2021)

is proposed an interesting proposal of a psychologi-

cally driven ontology for describing human behavior

based on character traits and states. The OWL on-

tology they propose highlights that taxonomic knowl-

edge representation is not sufﬁcient to fulﬁll the pur-

pose. In fact, the use of Machine Learning methods to

predict mental state rather than a Semantic Web Rea-

soner is adopted. The ITS framework we propose will

work by applying multi-strategy reasoning rather than

Machine Learning methods.

2.2 KEPLAIR

KEPLAIR is an ITS designed to act as a personalised

tutor that helps learners ﬁnd the tools, resources, ex-

periences, and connections they need to learn exactly

what they have chosen to learn. KEPLAIR’s artiﬁcial

intelligence engine will facilitates ﬁne-grained per-

sonalisation for each user, taking into proper account

the many aspects involved: learner’s cognitive level,

pre-existing knowledge about the topic, and preferred

physical and social environments, needs, background,

abilities, aims, interests, tastes, preferences, attitudes,

behaviours, motivations, expectations, context, and

community. KEPLAIR’s recommendations might in-

clude books, audio ﬁles, videos, online courses, mem-

bership associations, community resources, or even

other people. KEPLAIR can help students ﬁnd their

way both through predetermined curricula, and on

their own educational goals. The system will perva-

sively use symbolic AI to carry out its tasks. Dis-

tinguishing characteristics and contributions of KE-

PLAIR are the extensive set of features it proposes,

covering all aspects of the learning process. To en-

sure homogeneity and coordination among all of its

subsystems and components, a crucial role is played

by an ontology. It will act as a schema for the data that

informs all the internal representations and behavior,

so as to smoothly connect and orchestrate all the var-

ious functions and ensure both internal and external

interoperability. In particular, interoperability with

GraphBRAIN technology (detailed in the next sub-

section) is guaranteed since it has the same organiza-

tion of knowledge (Ferilli et al., 2022). KEPLAIR

was conceived with the aim of supporting students

independently of educational authorities, although it

can also be used to support teachers and education

managers. Since one of the PCTO objectives is the

KEOD 2022 - 14th International Conference on Knowledge Engineering and Ontology Development

254

involvement of learners in the deﬁnition of learning

path and in the choice of study materials, KEPLAIR

seems well suited for this purpose as we explain in the

proposed approach section.

2.3 GraphBRAIN

GraphBRAIN (Ferilli and Redavid, 2020a) is a

general-purpose tool that allows to design and col-

laboratively populate knowledge graphs, and provides

advanced solution for their fruition, consultation and

analysis. The functions provided by GraphBRAIN

bring to cooperation many different Artiﬁcial Intel-

ligence tasks, techniques, and approaches for im-

proving knowledge management and (personalized)

fruition by users, including: database technology, on-

tologies, data mining, machine learning, automated

reasoning, natural language processing, personaliza-

tion and recommendation, collaborative and social in-

teraction tools, and social network analysis. While

most of these items are investigated and exploited

separately in the state-of-the-art, the relevance of the

GraphBRAIN methodology is in their being really in-

tegrated, and not simply juxtaposed, so that each of

them takes directly or indirectly advantage from all

the others. This allows GraphBRAIN to ﬁnd relevant,

personalized, and non-trivial information, e.g., a so-

cial approach is used to build and integrate ontologies;

user models are used to guide data mining; ontologies

are used to guide database interaction and interface

generation; data mining is used to ﬁlter a manageable

and relevant portion of a huge graph on which car-

rying out automated reasoning, etc. GraphBRAIN’s

functions can be used on-line, interactively by end

users or delivered as Web services to other applica-

tions for obtaining selective and personalized access

to the stored knowledge. For the formal representa-

tion of emotional states only a taxonomic representa-

tion model may not be sufﬁcient, but more expressive

models are needed. For this reason, GraphBRAIN

was considered in this proposal. In fact, it supports

multiple formal knowledge representation languages

ranging from OWL (so it is possible to reuse existing

ontologies) to FOL (more ﬂexible for the representa-

tion of variable knowledge structures such as context-

and situation-dependent ones).

2.4 WoMan

WoMan (Workﬂow Manager) (Ferilli, 2014) is a

framework for workﬂow learning and management,

based on First-Order Logic representations. Its pro-

cess mining engine, applicable to activity logs coming

from actual process executions, is able to learn mod-

els involving concurrent, repeated, optional and dupli-

cate tasks in any combination, weighted elements. Its

full incrementality avoids the need for having all the

examples available from the beginning, still allowing

the learning to start from scratch. Correct models can

be learned using very few examples (in principle, any

set of examples including at least one representative

of each allowed process). It can also handle noise in a

very straightforward, intuitive way. Finally, the repre-

sentation language allows the description of not just

the ﬂow of events, but also the context in which the

activities take place, and hence the learning of com-

plex (and human- readable) pre- and post-conditions

for the workﬂow elements. A relevant issue in Pro-

cess Management in general, and in Process Mining

in particular, is to assess how well can a model pro-

vide hints about what is going on in the process exe-

cution, and what will happen next. Indeed, given an

intermediate status of a process execution, knowing

how the execution will proceed might allow the (hu-

man or automatic) supervisor to take suitable actions

that facilitate the next activities. The task of activity

prediction may be stated as follows: given a process

model and the current (partial) status of a new process

execution, guess which will be the next activity that

will take place in the execution (Ferilli et al., 2017).

WoMan models can be used for the monitoring and

supervision of processes and, when applicable, can be

translated into standard representations (Petri nets).

Both controlled and real-world experiments show that

WoMan outperforms existing process mining systems

in accuracy, effectiveness and efﬁciency. It ensures

quick, correct convergence towards the correct model,

using much less training examples than would be re-

quired by statistical techniques, even in the presence

of noise. WoMan is currently being wrapped in a Web

service that can be exploited by external applications

for learning, simulation and checking of workﬂows.

3 AN APPLICATION CASE

3.1 EU Recommendation on

Transversal Competences

The European Union has deﬁned transversal compe-

tences as those skills that enable citizens to act con-

sciously in a profoundly complex social context and

to meet the challenges posed by increasingly digi-

tised and interconnected organisational models. The

European Council (with the Recommendation of 22

May 2018) summarised the transversal competences

by specifying a comprehensive framework structured

according to the speciﬁc competence elements. This

Emotional States Management for an Advanced Intelligent Tutoring System

255

framework is organised according to four semantic ar-

eas and one in particular is related to emotional states:

• Personal, Social and Learning to Learn. Com-

petence refers to the ability to manage one’s own

learning, to lead a physically and mentally healthy

life, to create the right conditions to work well in a

group, to act in complex situations and to manage

interpersonal dynamics in an inclusive and con-

structive perspective.

Transversal competences are placed at the centre

of the learning pathway because they improve the

learner’s degree of awareness of his or her own per-

sonal growth. At the same time, they activate reﬂec-

tive and behavioural skills that are essential for mov-

ing around in social and work contexts; in fact, they

involve processes of thought and cognition, but also

of behaviour. They are key competences in the per-

spective of lifelong learning because they are char-

acterised by a high degree of transferability to differ-

ent tasks and environments, thus equipping the learner

with skills that enable him/her to improve the quality

of his/her own behaviour and to implement effective

strategies for the different contexts in which he/she

will be acting. Furthermore, it is important to con-

sider the importance of these soft skills also in a self-

orientative function: the student must be able to ob-

tain feed-back on his strategies and use them to re-

organise his ability to orientate himself in different

areas. In short, transversal competences enable the

student to enrich his personal assets with knowledge,

skills and attitudes that enable him to behave ade-

quately and effectively in the complexity of the sit-

uations in which he ﬁnds himself.

The peculiar nature of transversal competences

implies for the school an innovation in teaching

methodology, oriented towards strengthening the con-

nection between formal, informal and non-formal

contexts in which learning takes place. The emo-

tional and relational aspect is placed at the centre

of the educational process and becomes a substan-

tial element of lifelong learning. In the same way

as the teaching methodology, the monitoring of the

learning process, and thus the assessment tools, must

also be adapted to the characteristics of transversal

competences. Among other things, this also means

organising and prioritising individual and group in-

terviews, simulations and other active methodolo-

gies (role playing, project work, etc.) over ’tradi-

tional’ forms of assessment. Assessment, in fact, no

longer concerns only the goals and skills acquired, but

also the degree of awareness acquired by the student,

ﬁrst and foremost in knowing how to judge and en-

hance his or her abilities in terms of transversal skills.

Therefore, in line with the general teaching approach,

the activation and participation of the student is also

a central element for the monitoring and evaluation

system of the training pathway.

3.2 Transversal and Orientation

Competences

The EU recommendations have been implemented

in Italy through Transversal and orientation compe-

tences (denoted as PCTO) guidelines. The main pur-

pose of the PTCO is to make sure the student acquires

the functional competences for the study pathway un-

dertaken and the transversal competences aimed at

orientation in the world of work or at subsequent

higher education. This means covering the curricular,

the experiential and the orientation dimensions.

Starting from this fundamental premise, the PT-

COs can be develop with different organisational

forms, not only according to the course of study or

to the territorial speciﬁcity of the school, but also ac-

cording to the personal needs of each student. Person-

alisation of the study pathway is an essential aspect

because it allows the student to become aware and

self-directed in deﬁning his or her personal growth

project. This is why it is possible to develop differ-

ent types of PTCO within the same class group. Fur-

thermore, it must be considered that the possibility

of realising the pathway abroad is also envisaged, al-

ways as a function of an activity that is as coherent

and functional as possible in relation to the student’s

speciﬁc pathway. For this reason, the programme

presents multiple options with respect to the organisa-

tion with which the school can enable collaborations.

In addition to public and private bodies, third sec-

tor and entrepreneurial entities are becoming increas-

ingly important. Therefore, the design of a PTCO in-

stance must have ﬂexibility as a fundamental organi-

sational criterion, but within a well-deﬁned regulatory

framework.

3.3 PCTO Roles and Evaluation

Fundamental to the success of a PTCO is the role of

the disciplinary departments, whose task is to ensure

consistency with the three-year Education and Train-

ing Offer Plan as prescribed in Italy. However, it is

the Class Councils that design (alone or in collabo-

ration with the external body) the pathway, manage

the activities and carry out the ﬁnal assessment. In

fact, ﬁrst the Class Council selects the competences

for the class group, then each individual teacher must

identify (from among these selected competences) the

speciﬁc ones that he or she considers functional to his

or her teaching. The careful selection of the com-

KEOD 2022 - 14th International Conference on Knowledge Engineering and Ontology Development

256

petences to be developed has a fundamental impor-

tance. It must allow for student self-direction, in-

volving them already in the planning of activities and

stimulating their reﬂection and active participation.

Similarly, communication with families, documenta-

tion of all the stages of the pathway and sharing the

results of the experience are critical for the outcome

of the PTCO. Finally, it is essential that the school in-

stitution moves towards co-design involving the exter-

nal parties in deﬁning the objectives and educational

methods in that cases when the project is carried out

in collaboration with a third party.

Coordination between the parties involved is per-

formed by the tutor, who is appointed by the educa-

tional institution to perform certain functions that are

fundamental for the implementation of the pathway.

In addition to the coordination between the school in-

stitution, the third parties involved and the family, the

tutor constantly monitors the development of the ac-

tivities, provides assistance to the student and informs

the school institution of any critical issues. He/she is

a crucial ﬁgure because he/she plays a managerial and

supportive role, which fosters the creation of the right

context for achieving the planned goals. If provided

for in the PTCO project, the internal tutor may be sup-

ported by an external tutor selected by the host organ-

isation. This ﬁgure is the student’s referent within the

organisation where the training activity takes place,

but also acts as a link between it and the educational

institution. He/she is therefore expected to maintain

a constant liaison with the internal tutor. The interac-

tion between the two tutors, who must be selected on

the basis of appropriate training skills, is a key factor

for the success of the course.

The evaluation of the PTCO should assess the pro-

cess and the end result. In this way, it is not only the

objectives achieved that are assessed. Through the

structured observation of the entire process, the ac-

quisition of transversal competences is also assessed,

giving importance also to character and motivational

aspects.

The most frequently used tools for participative

observation are rubrics, diaries, digital portfolios and

observation sheets. The ﬁnal results, on the other

hand, are evaluated in several stages, ranging from

the identiﬁcation of objectives to the veriﬁcation of

the content learnt during the course. Obviously, the

observation of the entire process (carried out by the

tutors) inﬂuences the evaluation of the ﬁnal results,

which is, however, done by the teachers of the class

council and affects behaviour and the ﬁnal score.

Lastly, it should be emphasised that PTCO activities

must be included in the student’s curriculum, the doc-

ument that is attached to the ﬁnal diploma to certify

the skills acquired by the student along the course of

study.

4 PROPOSED APPROACH

4.1 Approach Steps

In this section, we report how the proposed frame-

work works identifying how the prototypes imple-

mentation will be used. In detail, three phases can

be identiﬁed:

• Uploading content into KEPLAIR. In this phase

the school institution uploads, if not pre-existent,

the content that it considers to be useful for the

speciﬁc transversal competences pertaining to the

institution itself. This content will enrich the

material contained in KEPLAIR and will remain

available to the entire community that will use it.

• Creation of the PCTO project. As prescribed by

the PCTO guidelines, a learning pathway that will

be assessed by the school institution must be cre-

ated. In this phase, the active involvement of the

student is explicitly stated in the EU recommen-

dations. The student will then use KEPLAIR in-

dependently obtaining the possible learning paths.

The material to be studied suggested in a learn-

ing path may or may not include the material up-

loaded in the previous phase by the institution.

In fact, KEPLAIR will determine to recommend

or not the institution contents by applying its AI

approaches on the whole material available in its

knowledge base. Finally, the student will dis-

cuss with the tutor the choice of pathway to be

submitted for approval by the educational institu-

tion from among those suggested by KEPLAIR.

The approved learning pathway will be seman-

tically formalised through a process speciﬁed in

the WoMan formalism. In this process there will

be the basic workﬂow represented by the chosen

learning pathway and some features chosen by the

institution useful to monitor learning progress in

accordance with PCTO guidelines. In particular,

these features will include those related to per-

sonal, social and learning to learn skills speciﬁed

in the section 1.

• Course delivery. During the delivery of the

content, physical or virtual sensors, part of the

WoMan process, are used to detect values useful

to establish the mood of the learners. These val-

ues are stored in the graph DB managed by Graph-

BRAIN. This tool is enabled to manage these data

semantics by means of an ontological representa-

Emotional States Management for an Advanced Intelligent Tutoring System

257

tion that allows either taxonomic (e.g., OWL) or

more generic (e.g., FOL) representations. As stip-

ulated in the PCTO guidelines, the learning path

is evaluated at precise points of the activities to be

performed and also takes into account the moods

of the learners useful for monitoring the learning

progress. Since the semantics of all this informa-

tion can be interpreted by GraphBRAIN, anoma-

lous situations can be detected applying the multi-

strategic reasoning (Ferilli and Redavid, 2020b).

In this case particular information will be com-

municated to KEPLAIR that will modify the sug-

gested learning path. As consequence, the ma-

terial that was suggested to the learners and ap-

proved by the teachers can be modiﬁed in order

to improve the learning outcome (e.g. if Graph-

BRAIN detects a certain discomfort in the student

since supplementary material in Maths has been

assigned than KEPLAIR suggests alternative sup-

plementary material). In addition, the supervi-

sion functionality offered by WoMan will make

it possible to assess whether the speciﬁc student

is correctly following the process formalised for

him/her in the previous phase. Through this mon-

itoring, corrective actions can be taken in accor-

dance with the tutor in order to have a better result

of the student’s learning performance.

5 CONCLUSIONS AND FUTURE

WORKS

The ITS meta-analytic review presented in (Kulik and

Fletcher, 2016) has proved their efﬁciency with re-

spect to other forms of tutoring. As the authors them-

selves indicate, it is not easy to determine what the

next generation of ITS will look like, but they will cer-

tainly be inﬂuenced by three factors: computer hard-

ware, software, networking, and cognitive science. In

this paper we have outlined a possible framework that

can also handle the moods represented by learning

processes formalised in the WoMan formalism. Fur-

thermore, through GraphBRAIN it will be possible

to handle different knowledge representations and ap-

ply multi-strategic reasoning in order to improve stu-

dents’ learning performance. In addition, by using

KEPLAIR we are able to cover one of the fundamen-

tal requirements of PCTO: empowering students by

allowing them to create proposals for learning paths.

As future work, a comparison with realities outside

the European community is desirable. In particular,

we have planned to test the platform in the Ukrainian

context with the support of the T.H. Shevchenko Na-

tional University ”Chernihiv Colehium”. On the one

hand, the sharing of EU recommendations can be

an opportunity to bring these two realities closer to-

gether; on the other hand, the current difﬁcult situa-

tion requires a massive use of online tools for school

education. Also if any e-learning has different goals

(Matviichuk et al., 2017), the deeper use of modern

technologies, namely elements of the artiﬁcial intelli-

gence (AI) proposed in this work, could be reveal new

modality to use an intelligent learning system (ITS).

ACKNOWLEDGEMENTS

This work is partially founded by the Apulia Region

through the ‘Research for Innovation—REFIN’ ini-

tiative [grant number 48B6D67E].

REFERENCES

Abaalkhail, R., Guthier, B., Alharthi, R., and El-Saddik, A.

(2018). Survey on ontologies for affective states and

their inﬂuences. Semantic Web, 9(4):441–458.

Bolock, A. E., Abdennadher, S., and Herbert, C. (2021). An

ontology-based framework for psychological monitor-

ing in education during the covid-19 pandemic. Fron-

tiers in Psychology, 12.

Ferilli, S. (2014). Woman: Logic-based workﬂow learn-

ing and management. IEEE Transactions on Systems,

Man, and Cybernetics: Systems, 44(6):744–756.

Ferilli, S., Esposito, F., Redavid, D., and Angelastro, S.

(2017). Extended process models for activity predic-

tion. In Kryszkiewicz, M., Appice, A.,

ezak, D., Ry-

binski, H., Skowron, A., and Ra

s, Z. W., editors, Foun-

dations of Intelligent Systems, pages 368–377, Cham.

Springer International Publishing.

Ferilli, S. and Redavid, D. (2020a). The graphbrain sys-

tem for knowledge graph management and advanced

fruition. In Helic, D., Leitner, G., Stettinger, M.,

Felfernig, A., and Ra

s, Z. W., editors, Foundations of

Intelligent Systems, pages 308–317, Cham. Springer

International Publishing.

Ferilli, S. and Redavid, D. (2020b). An ontology and knowl-

edge graph infrastructure for digital library knowledge

representation. In Ceci, M., Ferilli, S., and Poggi, A.,

editors, Digital Libraries: The Era of Big Data and

Data Science, pages 47–61, Cham. Springer Interna-

tional Publishing.

Ferilli, S., Redavid, D., Di Pierro, D., and Loop, L. (2022).

An ontology-driven architecture for intelligent tutor-

ing systems with an application to learning object rec-

ommendation. International Journal of Computer In-

formation Systems and Industrial Management Appli-

cations, 14:297–312.

Ferilli, S., Redavid, D., Pierro, D. D., and Loop, L. (2021).

Functionality and architecture for a platform for in-

dependent learners: KEPLAIR. In Abraham, A.,

KEOD 2022 - 14th International Conference on Knowledge Engineering and Ontology Development

258

Gandhi, N., Hanne, T., Hong, T., Rios, T. N., and

Ding, W., editors, Intelligent Systems Design and Ap-

plications - 21st International Conference on Intelli-

gent Systems Design and Applications (ISDA 2021)

Held During December 13-15, 2021, volume 418 of

Lecture Notes in Networks and Systems, pages 795–

805. Springer.

Imani, M. and Montazer, G. A. (2019). A survey of emo-

tion recognition methods with emphasis on e-learning

environments. Journal of Network and Computer Ap-

plications, 147:102423.

Kulik, J. A. and Fletcher, J. (2016). Effectiveness of intelli-

gent tutoring systems: a meta-analytic review. Review

of educational research, 86(1):42–78.

Matviichuk, L., Kukhar, L., and Nataliia, H. (2017). Exam-

ining factors of using information and communication

technologies for e-learning organization. Science Ed-

ucation, 26:68–73.

Yadegaridehkordi, E., Noor, N. F. B. M., Ayub, M. N. B.,

Affal, H. B., and Hussin, N. B. (2019). Affective com-

puting in education: A systematic review and future

research. Computers & Education, 142:103649.

Emotional States Management for an Advanced Intelligent Tutoring System

259