(IJACSA) International Journal of Advanced Computer Science and Applications,

Vol. 13, No. 8, 2022

Towards Personalized Adaptive Learning in

e-Learning Recommender Systems

Massra Sabeima Myriam Lamolle Mohamedade Farouk Nanne

CSIDS-University of Nouakchott LIASD, IUT de Montreuil-University Paris8 CSIDS-University of Nouakchott
LIASD, IUT de Montreuil-University Paris8 Montreuil, France Nouakchott, Mauritanie
Montreuil, France

Abstract—An adaptive e-learning scenario not only allows • choosing the right technologies and realization of that
people to remain motivated and engaged in the learning process, adaptation on a computer.
but it also helps them expand their awareness of the courses
they are interested in. e-Learning systems in recent years had Brusilovsky and Millan [3] on the other hand put focus
to adjust with the advancement of the educational situation. on the user modeling inside an adaptive system where the
Therefore many recommender systems have been presented to user information are a distinctive aspect to consider when
design and provide educational resources. However, some of the the system intervene. The interaction of the user should be
major aspects of the learning process have not been explored quite noticed with attention, when she/he searches, navigates; but
enough; for example, the adaptation to each learner. In learning,
and in a precise way in the context of the lifelong learning process,
also her/his interest, knowledge, background, learning style,
adaptability is necessary to provide adequate learning resources goals, etc. The priority should be given to the content suitable
and learning paths that suit the learners’ characteristics, skills, to what user interest in the most. User modeling featured-
etc. e-Learning systems should allow the learner to benefit the based or stereotype-based [3] should either way take into
most from the presented learning resources content taking into consideration the personal information of the individual. A
account her/his learning experience. The most relevant resources definition of adaptation is the reconfiguration of entities in
should be recommended matching her/his profile and knowledge order to adjust them to a certain request. It can be categorised
background not forgetting the learning goals she/he would like as the following according to [4]:
to achieve and the spare time she/he has in order to adjust the
learning session with her/his goals whether it is to acquire or • Machine Centred: In this case, the learning process is
reinforce a certain skill. This paper proposes a personalized e- guided by a series of actions from user and analyzed
learning system that recommends learning paths adapted to the by her/him.
users profile.
• User Centred: The learning resources (lessons) are
Keywords—e-Learning; adaptive learning; recommendation sys-
personalized by learners themselves as stats [4].
tem; ontology
Underneath these categories we find several kinds of adap-
tation [3], [4]. We mention:
I. I NTRODUCTION
• Adaptation of Content/Adaptive Evaluation: The con-
With the broad coverage of the internet, access to learning
tent of activities and resources are faced to dynamic
content through the web has become increasingly easy. A
change.
variety of educational systems such as MOOCs1 [1] have
emerged, with an essential mission that is provide educational • Adaptation of Visual Presentation: It represents
content, to learners willing to learn; yet the diversity of mainly the components of an interface and their prop-
people implies that each learner has her/his own particular erties, how and where they are displayed.
preferences, knowledge and competencies. In that perspective
adaptability was a major and essential criteria to add to e- • Adaptation of Learning Process: The learning process
learning systems providing learning resources, to make learn- is dynamically modified to the manner in which the
ing content suitable to learners. This adaptation takes a process courses contents are provided in suitable ways.
that is established in many levels. At the cognitive model level • Adaptive Information Filtering: The system takes care
as Ruiz et al. [2] propose, it have to go through the following of suitable information retrieval in order to give rele-
steps: vant results to user.
• to classify the user by choosing a suitable learning • Adaptive User Grouping: This allows a distant learn-
style; ers to collaborate and provide assistance in achieving
specific tasks.
• to present adaptation to system by developing good
techniques then conceive that adaptation to suits the However we could not talk about adaptation in a system
user’s preferences; without mentioning personalization which according to [5] is
included into a simple mechanism that need specific technolo-
1 Massive Open Online Courses gies to ensure accurate results. Adaptation inside a system
www.ijacsa.thesai.org 14 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 8, 2022

takes multiple parameters. The most important one, and re- a missing or unknown item of a dataset, and ii) the descriptive
gardless the technique used, is the user profile and information; (clustering) one which is based on grouping similar objects.
taking into consideration that the user interest and motivation The primary uses of clustering are to segment or categorize
are what will keep her/him continue learning. Therefore well (e.g., sorting customer data by age, occupation, and residence)
modeling her/his profile is essential. This profile is a personi- or to extract knowledge in an effort to find subsets of data
fication of different features of user. We note this profile can that are challenging to categorize. This method is about deter-
be built from two types. The first one constitutes a general mining a class for an element in a dataset. For instance, we
profile not specific to any user characteristic. The second can think of prediction as anticipating the appropriate course of
one is the developed version of the former. After extracting treatment for a certain disease in a specific individual. Whereas
the user information, a personalized profile is created to the grouping of patients based on their medical records can be
represent her/him. This study provides a critical overview of considered classification.4
previous research related to adaptive recommender systems
in e-learning field. Mainly we seek to answer the following George and Lal [7] show how ontology-based recom-
research questions: mender systems became an emergent research way in the e-
learning field. Those systems address most of issues found
• What aspect of adaptation should be enhanced and in e-learning recommender systems. Giving personalized rec-
why? ommendations to learners is one of the practical applications
of employing ontology-based recommender systems. Based on
• How is adaptation implemented in recommendation the learner’s interests, goals, etc., the recommendations that
systems? are given to them become precisely relevant. As a result,
• How to enhance it? the learner is encouraged to finish what she/he started. They
illustrated their point of view after a study on research papers
This paper is organised as follow: Section II will outline published during the last decade concerning the recommender
similar research work papers. Then in Section III we will un- systems in e-learning. They present extraction and modeling
derline common techniques and methods used for adaption in techniques used and compare existing recommender systems
adaptive e-learning systems. Section IV presents a comparison in e-learning in the scope of these techniques. From another
about different adaptive systems. Their advantages and draw- perspective Eke et al. [8] focused on user profiling methods,
backs are highlighted. Section V details our proposition, then and the challenges such as multi-dimensional representation,
in Section VI, we make a position from the current research privacy of user’s information, cold start problem for new
tendency in e-learning adaptive recommender systems and users, temporal behaviour of individuals, limitation of interest,
their techniques, which further consolidates our proposition in etc. They also discuss the most relevant solution for those
the previous section. Finally, Section VII concludes this paper challenges such as ontology representation and general purpose
and presents some perspectives. profile, and so on.
In [9], Nabizadeh et al. outline the personalization methods
II. A DAPTIVE E-L EARNING S YSTEMS and illustrate the challenges facing those methods, and how
to improve the existing personalization techniques. Zaoudi et
Many studies have been conducted in the field of adaptive
al. [10] present a critical research paper on existing approaches
e-learning systems. In this section, we present an overview
used in learning scenarios and adaptive e-learning situations.
of the research papers selected from the literature review. We
mention most relative ones from 2017 to 2021. This selection Then Javed et al. [11] present review of a widely used
was based on the relevance and the level of the adaptability in methods in recommender systems, context-based and content-
e-learning systems in terms of adaptive information filtering, based, and a hybrid method combining multiple methods in
adaptive user profile and adaptive users group. The process of order to benefit of the advantage of each method to cover the
selection was made as follows: among many articles found disadvantages of each one. Just recently, Raj and Ranumol [12]
in Google scholar, ResearchGate2 , DBLP3 while searching provide a review of research papers on a period of time from
for keywords such as “adaptation in e-learning systems”, 2015 to 2020, with critical study of adaptive recommender
“recommendation techniques”, “adaptative systems”, etc. 26 systems proposed comparing on one hand methods used in
papers were selected based on their relevance, their accuracy, those systems, from the hybrid methods, content or agent-
and the number of citations in others research works. Of these based, semantic web based, etc. On the other hand, they are
26 articles, we have selected seven (7) to be mentioned, based also comparing the attributes such as the user content rating,
on the year of publication and direct projection of their content learning style, knowledge level, etc.
on our research.
Table I summarises our comparison of recent research
Almmouhamadi et al [6] present a survey on the rising works, which we analyse in Section IV.
techniques of adaptation in educational adaptive systems. They
emphasize the two most used techniques of data mining in III. R ECOMMENDATION T ECHNIQUES
AI: (i) the predictive which is a prediction of the next tag
in general. By selecting a predictor variable or group of Adaptive recommendation systems can be divided into
variables, these techniques are applied to extract single or knowledge-based, content-based, user-based or based on hy-
multiple variables with predicted values; it is about predicting brid approaches. They can be categorised according to what the
adaptation is based on and on the recommendation techniques
2 https://www.researchgate.net/
3 https://dblp.uni-trier.de/ 4 see classification-and-prediction-methods-in-data-mining

www.ijacsa.thesai.org 15 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 8, 2022

(a) Instance parent1, individual of class Human

(b) After inferences

Fig. 1. Example of Inferred Knowledge by the OWL Reasoner HermiT.

used. Following those used techniques the adaptation can be Fig. 1(b) displays, after starting the reasoner
based on: HermiT, new knowledge that has been inferred
about Instance Child1 which belongs to class
• User’s Profile: this method takes into account the (noted type in Fig. 1) Human, represented in
characteristics of the user defined by her/his intrinsic description logic by Human(Instance Child1)
characteristics, her/his preferences for the presentation and the new object property assertion
of pedagogical resources to be recommended (text, isChild(Instance Child1, Instance parent1.
audio, video, etc.), and the experience of other users
with similar profiles. To do this, several techniques • Knowledge-Based: this approach with slight similarity
have been implemented to model user profile. This al- with the user profile based approach, in representing
lows a learner model to be designed which, according knowledge. It helps making recommendation by ex-
to [13], is the representation of specific characteristics tracting information. In general, a reasoning system
of a learner that may be relevant for a personalized is behind that decision making, after having well
interaction. Managing users’ profile allows the user represented knowledge.
learning style to be predicted. Most of them are based item Content-Based: which is based on the content
on the widely used learning model “The Felder and of the started themes. The evaluation of the content
Silverman Learning style Model” (FSLSM) [14]. It is is done in an explicit way by the attribution of
worth noting that the learner model is not intended to notes directly to the documents which represent the
be a representation of the learner’s mental state but contents of the topics, or in an implicit way when the
rather of the learner’s characteristics such as personal system estimates through user interactions the degree
information (age, gender, country, native language, of relevance of a document [18].
etc.), cognitive traits, knowledge and skill levels, pre-
ferred learning styles, and personal preferences, such • Collaborative Filtering: which is a widely used
as cultural background, format of learning resources method that consists in projecting the preferences of
(text, audio, video, etc.), preferred language, etc. Mu- an individual to a group of similar users. In other
nassar and Ali [15] and Aissaoui and Oughdir [16] words, the recommendation is made on the basis of
propose a framework based on user profile modeling what our neighbors (users with similar profiles) have
using ontologies, which represent the terminology appreciated [19];
(TBox) and assertions (ABox) such as instances of • Social-Based: basically these methods can be used to
concept. From a knowledge base, a reasoner checks enhance an already existing system, by using social
the consistency of the model and infer new knowledge network to create similar groups [20]. One might
depending of description logic level used [17]. assume that users who are friends on social networks
For example, let us an ontology O (for didactic can have a common interest, or even one user can be
purposes): interested in a resources because her/his friends were
Human(Instance parent1) or are interested in taking it. It would be interesting
isP arent(Instance parent1, Instance Child1) to detect the influencers. User activity on social-media
Human.isChild ≡ Human.isP arent− also in recent years formed a good source for recom-
We can see below in Fig. 1 how the reasoner HermiT5 mendation, the time she/he spends watching a video
inferred new knowledge. can give an idea on languages she/he understands,
Fig. 1(a) shows an individual Instance parent1 of and the subjects that interest her/him. Furthermore the
a class Human with a role (i.e. Object Property content she/he likes and comments or shares also are
assertion in the editor Protégé6 ) isP arent, which is considerable source. Her/his geolocation history can
the inverse role of isChild. also be known through her/his publications and the
5 www.hermit-reasoner.com/ location she/he visits. A variety of information can be
6 protege.stanford.edu/ extracted through social networks.
www.ijacsa.thesai.org 16 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 8, 2022

TABLE I. C OMPARISON OF A DAPTIVE R ECOMMENDER S YSTEMS IN E-L EARNING

Reference KB CB UB MU
Sarwar et al. [22] X - X hybrid techniques + ontologies
Agbonifo et al. [23] - X X collaborative filtrage + ontologies
Vagale et al. [24] - X X extraction form user model
Boussakssou et al. [25] X X - Q-learning
Shi et l. [26] X X - A designed knowledge graph +Bloom’s taxonomy
Azzi et al. [27] X X - Automatic prediction of Learning style + Fuzzy C-means
El Fazazi et al. [28] X X X MAS+Q-Learning algorithm

• Hybrid Methods: they combine two or more ap- Content modeling of learning objects also is another part
proaches of the previous types of recommendation that should be considered as important as the profile modeling.
techniques [21]. For example, it can be based on Moreover it is worth noting the low number of pedagogical
user characteristics by modeling the learner’s profile resources used in evaluation when seeking for users ratings
in the first step and, in the second step, recommend- about a course.
ing resources adaptively with respect to the profile.
Usually the combination of these techniques aims to On the other hand the hybrid techniques implemented in
overcome the drawbacks such as the sparsity issue those systems whom the semantic part for modeling content
which is due to lack of user rating. Users are reluctant and profiles candidate them to be technically efficient as they
to give feedback on items they have tested. In addition are using latest technologies tendencies.
to the sparcity, we mention the cold start problem
faced by new users or new item. This issue shows up, Table II gathers previously mentioned systems and lists the
when there is not a review/ratings of an item, making types of adaptation presented in the proposed work. We notice
it unrecommendable despite its importance and rele- that only Boussakssou et al. [25] integrate an adaptation based
vancy. The same applies on new users. The difficulty on user action in their proposed work. Whereas [22], [24] de-
of making recommendations based on a user’s profile scribed a group based adaptation and then a content adaptation
increases for new users with new profiles. in [24]. In [28] along with [24] propose an adaptation model
Hybride methodes benefits of the advantages of each that assures course adaptation. This adaptation relies on certain
techniques used. However, some limitations still per- characteristics of the user such as background Knowledge and
sist and even new challenges appear in hybrid ap- learning style using Q-learning in El Fazazi et al.s’ works [28].
proaches.
These comparison criteria are selected on the basis of
the definition and the different types of adaptation mentioned
IV. C OMPARISON OF A DAPTIVE R ECOMMENDER S YSTEM above.
In this section, we compare adaptive recommender sys-
tems proposed firstly in terms of adaptability techniques
used: Knowledge-based (KB), Content-Based (CB), User- V. M ETHODOLOGY
based (UB) and Method-Used (MU);
Following our findings on the advantages and disadvan-
and secondly in terms of adaptation itself, based on the tages of current adaptive recommendar systems, we propose a
two types mentioned in above Section I new architecture for piloting and customizing adaptive learning
Table I gathers recently proposed adaptive e-learning sys- paths, while taking into account the users’ profile, the training
tems using different methods. We notice that [22], [23] are domain and the available educational resources, and adding
generally using ontology to model content or user profile synchronization in the collaborative mode between learners
in their work along side with machine learning techniques, wishing to work in collaboration. This system is based on
whereas [26], [24], [25] used knowledge designed graphs to ontologies and a multi-agent system responsible of managing
represent the user model, or Q-learning [29] with machine events that occur inside the system. Reasoning on ontologies
learning techniques. allows to make tacit information explicit. Among other things,
this allows for a better personalization of learning paths. On
On the other hand, Azzi et al. [27] are more focused on the other hand, multi-agent systems have shown their great
users learning style. They proposes an approach that predicts capacity to orchestrate in real time a set of agents.
the user learning style and stores then the collected data. We
notice that most of the adaptation techniques used belong to
two main ranges, to ontology for modeling and representation,
and to machine learning methods. Also other researches tend A. Adaptive and Collaborative Learning Piloting Architecture
to combine two at least methods in order to develop an hybrid
technique. This architecture is composed of a multi-agent system
(MAS), which contains an agent manager representing the
These systems mainly evaluate the user performance and entry point of the main MAPE-K loop of the platform (cf.
make recommendation based on collective preferences like Fig. 2). This Agent analyzes the requests, processes them
Agbonifo and Akinsete [23] in their work experiments. This and manages the communication between the recommendation
method is not the most efficient way due to the lack of agents (RA), responsible for managing the creation of learning
specification in those rating (based on what and by who). paths and the recommendation of educational resources.
www.ijacsa.thesai.org 17 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 8, 2022

TABLE II. C OMPARISON IN T ERM OF A DAPTATION

Reference User Based Machine Based
Sarwar et al. [22] - X
Agbonifo et al. [23] - -
Vagale et al. [24] - X
Boussakssou et al. [25] X -
Shi et al. [26] - -
Azzi et al. [27] - -
El Fazazi et al. [28] - X

describes learners, their specific information, their initial or

acquired skills and the personalized learning paths already
completed or ongoing.

Fig. 2. SPACe-L Architecture.

Fig. 4. Partial View of Training Ontology (TO).

B. Knowledge Base
The core of the platform is its knowledge representation
The ontology of the training domains (TO) describes the
by a network of ontologies describing three ontological mod-
competences to be acquired, the learning objects and the peda-
els representing users profiles, training domains and video
gogical resources that can be used in the different pedagogical
resources. This modularity is intended to facilitate the interop-
units (Fig. 4).
erability with other ontologies in the respect of the FAIR prin-
ciples7 . For example, the user representation (cf. Fig. 3) allows
from the ontology FOAF data to be integrated. For the training C. Multi-Agent System
domain, we can integrate ontologies describing competences
like the ontology COMP2 proposed by G. Paquette [30]. The multi-agent system (MAS) manages the events that
occur in the system, under the supervision of the Agent
manager who analyzes the requests received and manages
the situation according to its nature. It plays a double role
according to the learning situation, either individual through
a recommendation agent, or collaborative. It then manages a
network of recommendation agents for the synchronization
of learners. The recommendation agent (RA) is associated
to one user or group of users in collaborative situation. The
recommendation agent will take care of the generation of the
personalized path in the form of a graph by associating relevant
pedagogical resources to each node of the graph.
The graph generation is done according to the user profile
and the recommendation of the pedagogical resources accord-
ing to several criteria including the learners’ preferences, the
duration of the session, but also the qualitative evaluations of
Fig. 3. Partial View of User Profile Ontology (UPO). the other learners, trainers or experts. The objective here is to
maximize a fitness function (cf. Equation 1) which allows to
dynamically generate a personalized learning path according
The partial view of UPO shown in Fig. 3 contains the to:
personal information and preferences of the users. It mainly Xn
F = Wi Ci (1)
7 https://www.go-fair.org/fair-principles/
i=1

www.ijacsa.thesai.org 18 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 8, 2022

Where Wi is the weight that defines the importance of can be satisfied by multiple uses of semantic web. This latter
Ci and n is the number of selected criteria of a pedagogical one is Non-linear as it allows user to describe the situation
resource. that she/he is currently in; for example the purpose of her/his
learning, and the knowledge acquired. The Semantic Web is
also interactive that agent can use commonly agreed service
VI. D ISCUSSION
language, enabling collaboration between them. Despite the
Based on the observation form the study done in this learning resources being distributed on the web, they are linked
paper, it is obvious that adaptation in a recommender sys- to one or more commonly accepted ontologies in the scope of
tems is essential in order to provide the learner what suits semantic web (cf. LOD). Learning materials are distributed on
her/him. Therefore, machine learning, ontological and hybrid the web, but they are linked to commonly agreed ontologie(s).
techniques have been applied in different propositions. We This allows to build a course that is unique to the user by using
mention in particular the machine learning collaborative fil- semantic querying to find relevant subjects of interest [35].
tering technique for its frequent use in those systems. It is Application of semantic web can create a responsive learning
reasonable to understand it is widely implicated in most of environment, a personalized learning materials where user
adaptive e-learning systems since it is a new form of the most only receives what suits her/him, and as much as decentralise
traditional method of recommendation (recommendation based content possible.
on personal user experience). In addition, collaborative filtering
system features include recommending an item by classing a
list of object based on whether it might be interesting to the VII. C ONCLUSIONS AND F UTURE W ORKS
user. They include also predicting for a specific item and its Adaptation in e-learning systems represents a trending
rating by a user [31]. research area. In this paper, we presented different adaptive e-
It is required, however, to pay attention to some of the learning systems representative of different categories. Several
challenges that can and have arisen; such as users rating to a methods were experimented and compared, yet the existing
certain pedagogical object (courses, learning object, learning methods have both benefits and drawbacks. The conflict of
path, pedagogical resources, etc.). The integrity of that rating which one is more effective is still. Mainly the adaptation
cannot be measured in reality, without exposing publicly the inside an e-learning environment is user centered even though
interest of the user or her/his personal information, that leads many researches use the content-based method. Others tend to
to another privacy problem. In addition to these problems, predict the learning style of the learner, or extract knowledge
we include the unequal number of users and votes on object. from user interaction and navigation history; while others
George and Lal [7] have pointed out in their research works lean to use techniques like collaborative filtering and machine
that the number of users is higher than the number of votes. learning methods. Except for the fact that user profile modeling
remains the main axis to highly adapt content and the learner’s
Content-based method relies on the interaction of the user need and interest. This being said modeling user is not an easy
and data collection after. Therefore item description is as task to achieve, nor extraction her information by tracing her
much as important as the user behaviour, seeing that the interactions through the web. Modeling user profile extends
recommendation is established based on that. The steps of representing her interests, competences, expectations of the
content-based recommendation techniques are as follows: (i) at course and goals. It might reach her mental state at the learning
the start, item description is stored after analyse, to determine session and after. We highlighted that modelling these new
the preferences of a user regarding this item for future use; (ii) criteria implies a high complexity level in the adaptive process
then a comparison mechanism is done between user profiles inside an e-learning system. That said the Semantic Web in a
and attributes of these items to sort only related items with sim- side is one sophisticated way to model a profile through the
ilarities with that profile. That said, it still represents multiple use of ontology. This research field can be highly explored
drawbacks. Let us mentioning user preferences and interest that and be employed to improve the current state of adaptive
change and that affect directly the recommendation. Another e-learning systems, especially the collaborative learning type
issue is the privacy previously mentioned of the user. In order which represents an important type of learning and increase
for these methods to have accurate recommendation a large learners motivation to reach new competences or reinforce
and precise amount of information must be extracted, and that competences.
might expose the user privacy policy. Synonymy is also another
issue represented by the fact that some of items can have However some of the relevant questions in that regard
very close description but they still different, which lead to still exist such as how can the recommendation systems be
erroneous recommendation [32]. improved? And in a more specific manner how is the adequate
learning path recommended? How can one be sure we are
Thus, Semantic Web is a research field existing since actually getting the right pedagogical resources? All these
the late 90s, especially ontologies, which is promising due questions concerns individual learning situations, it remains
to its ability of sharing, reusing and inferring knowledge those regarding the collaborative situation mentioned earlier,
including through Linked Open Data (LOD), and its level of how will the synchronisation between learner be established?
interoperability. Moreover, it is a good candidate to the FAIR Even if established how will the adaptation be maintained?
principles (Findable, Accessible, Interoperable, Reusable) [33], How can we keep learners interested and motivated to finish
[34]. They have also managed to address most of these the training and benefit it the most? How can we integrate
problems. There is no uniform model for the learner profile citizen science in the scope of collaborative learning? These
or structured material in e-learning, which makes ontology research questions are important to analyse and to focus on.
even more relevant [7]. In addition, e-learning requirement In this paper we proposed a recommendation system based
www.ijacsa.thesai.org 19 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 8, 2022

on Semantic Web for knowledge representation and multi- [16] O. E. Aissaoui and L. Oughdir, “A learning style-based ontology
agent system that manages the different events in the system. matching to enhance learning resources recommendation,” in 2020 1st
It seems to us that is a good way to answer of the above ques- International Conference on Innovative Research in Applied Science,
Engineering and Technology (IRASET), 2020, pp. 1–7.
tions. However there are still several points to improve in order [17] F. Baader, “Description logics,” in Reasoning Web: Semantic Technolo-
to obtain an advanced adaptive system. starting with enriching gies for Information Systems,5th International Summer School 2009,
our ontology network with ontologies coming from some ser. Lecture Notes in Computer Science. Springer–Verlag, 2009, vol.
standards or norms of the educational sciences, and on other 5689, pp. 1–39.
side improving response times for learning path and learning [18] S. S. Khanal, P. Prasad, A. Alsadoon, and A. Maag, “A systematic re-
resource recommendations during users synchronization from view: machine learning based recommendation systems for e-learning,”
Education and Information Technologies, vol. 25, no. 4, pp. 2635–2664,
what it concerns the multi-agent system performance. 2020.
[19] M. Ramdane et al., “Les systèmes multi-agents dynamiquement adapt-
R EFERENCES ables,” Ph.D. dissertation, Université Mentouri Constantine, 2017.
[1] M. Zhu, A. R. Sari, and M. M. Lee, “A comprehensive systematic [20] J. K. Tarus, Z. Niu, and G. Mustafa, “Knowledge-based recommenda-
review of mooc research: Research techniques, topics, and trends from tion: a review of ontology-based recommender systems for e-learning,”
2009 to 2019,” Educational Technology Research and Development, Artificial intelligence review, vol. 50, no. 1, pp. 21–48, 2018.
vol. 68, no. 4, pp. 1685–1710, 2020. [21] B. Rawat, J. K. Samriya, N. Pandey, and S. C. Wariyal, “A comprehen-
[2] M. d. P. P. Ruiz, M. J. F. Dı́az, F. O. Soler, and J. R. P. Pérez, sive study on recommendation systems their issues and future research
“Adaptation in current e-learning systems,” Computer Standards & direction in e-learning domain,” Materials Today: Proceedings, 2020.
Interfaces, vol. 30, no. 1-2, pp. 62–70, 2008. [22] S. Sarwar, Z. U. Qayyum, R. Garcı́a-Castro, M. Safyan, and R. F.
[3] P. Brusilovsky and E. Millán, “User models for adaptive hypermedia Munir, “Ontology based e-learning framework: A personalized, adaptive
and adaptive educational systems,” in The adaptive web. Springer, and context aware model,” Multimedia Tools and Applications, vol. 78,
2007, pp. 3–53. no. 24, pp. 34 745–34 771, sep 2019.
[4] A. Klašnja-Milićević, B. Vesin, M. Ivanović, Z. Budimac, and L. C. [23] O. C. Agbonifo and M. Akinsete, “Development of an ontology-based
Jain, “Personalization and adaptation in e-learning systems,” in E- personalised e-learning recommender system,” International Journal of
learning systems. Springer, 2017, pp. 21–25. Computer (IJC), vol. 38, no. 1, pp. 102–112, 2020.
[5] S. LaCour, “The future of integration, personalization, and eportfolio [24] V. Vagale, L. Niedrite, and S. Ignatjeva, “Implementation of personal-
technologies,” Innovate: Journal of Online Education, vol. 1, no. 4, ized adaptive e-learning system,” Baltic Journal of Modern Computing,
2005. vol. 8, no. 2, pp. 293–310, 2020.
[6] K. Almohammadi, H. Hagras, D. Alghazzawi, and G. Aldabbagh, [25] M. Boussakssou, B. Hssina, and M. Erittali, “Towards an adaptive e-
“A survey of artificial intelligence techniques employed for adaptive learning system based on q-learning algorithm,” Procedia Computer
educational systems within e-learning platforms,” Journal of Artificial Science, vol. 170, pp. 1198–1203, 2020.
Intelligence and Soft Computing Research, vol. 7, no. 1, pp. 47–64, [26] D. Shi, T. Wang, H. Xing, and H. Xu, “A learning path recommendation
2017. model based on a multidimensional knowledge graph framework for e-
[7] G. George and A. M. Lal, “Review of ontology-based recommender learning,” Knowledge-Based Systems, vol. 195, p. 105618, 2020.
systems in e-learning,” Computers & Education, vol. 142, p. 103642, [27] I. Azzi, A. Jeghal, A. Radouane, A. Yahyaouy, and H. Tairi, “A robust
2019. classification to predict learning styles in adaptive e-learning systems,”
[8] C. I. Eke, A. A. Norman, L. Shuib, and H. F. Nweke, “A survey of Education and Information Technologies, vol. 25, no. 1, pp. 437–448,
user profiling: State-of-the-art, challenges, and solutions,” IEEE Access, 2020.
vol. 7, pp. 144 907–144 924, 2019. [28] H. El Fazazi, M. Elgarej, M. Qbadou, and K. Mansouri, “Design
[9] A. H. Nabizadeh, J. P. Leal, H. N. Rafsanjani, and R. R. Shah, “Learning of an adaptive e-learning system based on multi-agent approach and
path personalization and recommendation methods: A survey of the reinforcement learning,” Engineering, Technology & Applied Science
state-of-the-art,” Expert Systems with Applications, vol. 159, p. 113596, Research, vol. 11, no. 1, pp. 6637–6644, 2021.
2020. [29] C. J. Watkins and P. Dayan, “Q-learning,” Machine learning, vol. 8, no.
[10] M. Zaoudi and H. Belhadaoui, “Adaptive e-learning: Adaptation of 3-4, pp. 279–292, 1992.
content according to the continuous evolution of the learner during [30] G. Paquette, O. Marino, and R. Bejaoui, “A new competency ontology
his training,” in Proceedings of the 3rd International Conference on for learning environments personalization,” Smart Learning Environ-
Networking, Information Systems & Security, 2020, pp. 1–6. ments, vol. 8, no. 1, pp. 1–23, 2021.
[11] U. Javed, K. Shaukat, I. A. Hameed, F. Iqbal, T. M. Alam, and [31] J. B. Schafer, D. Frankowski, J. Herlocker, and S. Sen, “Collaborative
S. Luo, “A review of content-based and context-based recommendation filtering recommender systems,” in The adaptive web. Springer, 2007,
systems,” International Journal of Emerging Technologies in Learning pp. 291–324.
(iJET), vol. 16, no. 3, pp. 274–306, 2021.
[32] B. Patel, P. Desai, and U. Panchal, “Methods of recommender system: A
[12] N. S. Raj and V. Renumol, “A systematic literature review on adaptive review,” in 2017 international conference on innovations in information,
content recommenders in personalized learning environments from 2015 embedded and communication systems (ICIIECS). IEEE, 2017, pp. 1–
to 2020,” Journal of Computers in Education, pp. 1–36, 2021. 4.
[13] A. E. Labib, J. H. Canós, and M. C. Penadés, “On the way to [33] A.-L. Lamprecht, L. Garcia, M. Kuzak, C. Martinez, R. Arcila, E. Mar-
learning style models integration: a learner’s characteristics ontology,” tin Del Pico, V. Dominguez Del Angel, S. Van De Sandt, J. Ison, P. A.
Computers in Human Behavior, vol. 73, pp. 433–445, 2017. Martinez et al., “Towards fair principles for research software,” Data
[14] A. Adetunji and A. Ademola, “A proposed architectural model for an Science, vol. 3, no. 1, pp. 37–59, 2020.
automatic adaptive e-learning system based on users learning style,” [34] F. Garbuglia, B. Saenen, V. Gaillard, and C. Engelhardt, “D7.5
IJACSA) International Journal of Advanced Computer Science and good practices in fair competence education,” Dec. 2021. [Online].
Applications, vol. 5, no. 4, pp. 101–128, 2014. Available: https://doi.org/10.5281/zenodo.5785253
[15] W. A. Munassar and A. F. Ali, “Semantic web technology and ontol- [35] B. Dutta, “Semantic web based e-learning,” Documentation Research
ogy for e-learning environment,” Egyptian Computer Science Journal, and Training Centre Indian Statistical Institute, Bangalore, 2006.
vol. 43, no. 2, pp. 88–100, 2019.

www.ijacsa.thesai.org 20 | P a g e
 © 2022. This work is licensed under
https://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding
the ProQuest Terms and Conditions, you may use this content in accordance
with the terms of the License.