The Use of Self-Regulation of Learning in Recommender Systems:

State-of-the-Art and Research Opportunities

Alana Viana Borges da Silva Neo

1 a

, Jos

e Ant

ao Beltr

ao Moura

1 b

Joseana Mac

edo Fechine R

egis de Ara

ujo

1 c

, Giseldo da Silva Neo

2 d

and Olival de Gusm

ao Freitas J

unior

3 e

Center for Electrical and Computer Engineering, Federal University of Campina Grande, Campina Grande, Brazil

Campus Vic¸osa, Federal Institute of Alagoas, Vic¸osa, Brazil

Institute of Computing, Federal University of Alagoas, Macei

o, Brazil

Keywords:

Online Education, Self-Regulated Learning, Virtual Learning Environment.

Abstract:

Self-regulated learning is deﬁned as the degree to which students are metacognitive, motivationally, and be-

haviorally active participants in their learning. Learning can be inﬂuenced and improved to achieve successful

academic results. This article reviews the literature to analyze and compare learning self-regulation strategies

to recommend learning objects in the context of a Virtual Learning Environment. The results serve to map

the state of the art, main approaches, and characterizations of the topic of recommendation systems that use

self-regulated learning strategies to support students’ academic performance; and to identify promising oppor-

tunities for future research on the topic.

1 INTRODUCTION

Online education has brought new education oppor-

tunities, but it has also brought many challenges for

students, such as deciding what, when, how, and for

how long to learn (Cerezo et al., 2020). When stu-

dents learn in an online environment, they can hardly

regulate their learning, thus failing to achieve objec-

tives (Hidayah et al., 2018). It is necessary to sup-

port them to have autonomy in their learning (Pierrot

et al., 2021), and several studies have been proposed

to help them plan tasks and monitor their performance

(Afzaal et al., 2021).

Previous studies have shown that a lack of self-

regulated learning (SRL) skills can be a major fac-

tor leading to failure of students and dropout from

courses (Afzaal et al., 2021). In this sense, for on-

line learning to be successful, students need to possess

these SRL skills (Wang et al., 2021). Self-regulated

students are aware of their learning process and can

take an active role in adapting to different learning

https://orcid.org/0009-0000-1910-1598

https://orcid.org/0000-0002-6393-5722

https://orcid.org/0000-0002-8825-2581

https://orcid.org/0000-0001-5574-9260

https://orcid.org/0000-0003-4418-8386

environments (Leite et al., 2022). Given this, it is es-

sential to investigate which SRL strategies are most

effective, and which can be recommended, aiming to

help students improve their academic performance.

The purpose of this study is to analyze and com-

pare strategies on the topic of SRL strategies for

recommending learning objects in Recommendation

Systems (RS), in the context of a Virtual Learning

Environment (VLE). In turn, analysis and compari-

son allow one to glimpse the state-of-the-art and iden-

tify research opportunities on the topic. For that, the

article addresses ﬁve research questions (RQs) as pre-

sented in Table 1. RQ1 to RQ5 are answered by works

on the topic that have been reported in the recent liter-

ature and the answer to RQ6 is extracted from the an-

swers to RQ1-5. The study is carried out using a sys-

tematic literature review spanning back over 5 years

(from 2018 to mid-2023).

The main contributions of this study are (i) an up-

dated analysis of recent works on the topic that can

serve as support for guiding educators, IT profes-

sionals, and researchers interested in understanding,

building, or applying SLR-based RS to VLEs; (ii) the

provision of a synthesized body of knowledge for fu-

ture reference and research.

The remainder of the article is divided into 5 sec-

tions. Section 2 provides a brief overview of the con-

Neo, A., Moura, J., Régis de Araújo, J., Neo, G. and Freitas Júnior, O.

The Use of Self-Regulation of Learning in Recommender Systems: State-of-the-Art and Research Opportunities.

DOI: 10.5220/0012619400003693

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

In Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024) - Volume 2, pages 525-532

ISBN: 978-989-758-697-2; ISSN: 2184-5026

525

Table 1: Research Questions for the Systematic Review on

the Topic of Learning Self-Regulation in RS for VLE.

ID Research questions

RQ1 What is the impact of student self-regulation and interaction

dynamics on recommendation in VLEs?

RQ2 Which SRL strategies are carried out by students in a VLE?

RQ3 How does a student’s current posture inﬂuence their ability to

self-regulate their learning in a VLE?

RQ4 How to measure SRL?

RQ5 Which datasets are used in the research?

RQ6 Which are promising opportunities for further research on the

topic?

cepts of VLEs, RS, and SRL to facilitate the reading

of the following sections. Next, Section 3 will present

the methodology adopted in the research. Section 4

analyzes the ﬁndings regarding RQ1-5. Section 5

explores the research gaps in the ﬁndings (RQ6).

The said gaps, if narrowed, could further enhance

the topic and cover its applications more comprehen-

sively. Section 6 brings ﬁnal considerations.

2 THEORETICAL FOUNDATION

2.1 Virtual Learning Environment

The Virtual Learning Environment (VLE) is a tech-

nological educational resource in which the learning

process depends entirely on the use of a computer-

ized environment and/or online resources (Al-Obaydi,

2020). In VLE, students carry out educational activi-

ties, answer questionnaires, watch video classes, and

study reading materials (Al-Obaydi, 2020).

Technological educational resources, such as

VLE, can be a positive factor in the development

of SRL processes, enabling students to focus more

on the proposed activities (Lima and Silva, 2010).

Various VLEs have emerged over time to support

SRL by offering personalized instructions or feed-

back to students without teacher intervention (Wang

et al., 2022). Some studies developed VLEs com-

plemented with metacognitive supports, allowing stu-

dents to implement important metacognitive strate-

gies and demonstrated learning gains for students who

used the VLE (Hidayah et al., 2018; Odilinye and

Popowich, 2020).

Recommendations in a VLE provide additional

beneﬁts to students who follow them, improving their

motivation and performance (Takami et al., 2022).

The recommendation can be more useful because it

can use the students’ actions and interactions, ges-

tures and mouse clicks, learning patterns and pro-

cesses, reﬂecting students’ cognitive and metacogni-

tive events captured in VLEs (Cerezo et al., 2020).

2.2 Recommendation Systems

Recommender systems (RS) have gained popularity

in the educational ﬁeld, offering different types of

recommendations for students, teachers, and schools;

identifying interesting learning materials from a large

set of resources, and reducing information overload

by recommending the right content at the right time

and in the right format for the learner (Odilinye and

Popowich, 2020). These recommendations are impor-

tant for the learning process, allowing teachers and

students to ﬁnd content appropriately, according to

their proﬁle and needs (Brito et al., 2014; Dwivedi

and Roshni, 2017; Obeid et al., 2018).

Some RS are designed to support SRL by pro-

viding recommendations on demand or automati-

cally when certain conditions are met (Odilinye and

Popowich, 2020). For a self-regulated RS, personal-

ization of recommendations through student model-

ing is necessary (Hidayah et al., 2018). Personalized

learning recommendations are necessary to meet each

student’s speciﬁc learning needs and preferences and

improve the learning experience. Each learner has

individual needs and speciﬁc requirements, and the

learner model is used to capture information about

learner characteristics such as learning objectives,

learning style, prior knowledge, and more (Odilinye

and Popowich, 2020).

2.3 Self-Regulation of Learning

Self-Regulation of Learning (SRL) is deﬁned as the

degree to which students are metacognitive partic-

ipants (students’ ability to establish plans, sched-

ules, or goals to monitor or evaluate their learn-

ing progress), motivational (students who are self-

motivated and willing to take responsibility for their

successes or failures), and behaviorally active in their

learning. For learning to be effective, students need

to intentionally activate, sustain, and adjust their cog-

nition, affect, and behavior to achieve their learning

goals (Kuo et al., 2014; Wang et al., 2021).

A self-regulated learner is a student who ap-

proaches educational tasks with conﬁdence, dili-

gence, and resourcefulness. Therefore, self-regulated

students can evaluate their learning strategies and

choose their skills and areas of weakness, as they can

modify their learning strategies to achieve the desired

academic result (McLellan and Jackson, 2017; Wang

et al., 2022).

Choosing and monitoring learning strategies are

key factors in the student’s learning process (Hi-

CSEDU 2024 - 16th International Conference on Computer Supported Education

526

dayah et al., 2018). Proposals for SRL include

models that consider the regulation of affect, be-

havior, and cognition, recognizing the importance of

emotional management (Boruchovitch, 2014; Ben-

Eliyahu and Linnenbrink-Garcia, 2015). Studies in-

dicate that students’ academic performance depends

on several factors, including self-regulation processes

that contribute to motivation and academic learning

(Ben-Eliyahu and Linnenbrink-Garcia, 2015; Soares,

2018).

There are some ways to measure SRL, Pintrich

et al. (1991) developed the MSLQ metric scale -

Motivated Strategies for Learning Questionnaire (Pin-

trich et al., 1991; Polydoro and Azzi, 2009). MSLQ

uses 81 items to assess students’ motivational orien-

tation and learning strategies in a speciﬁc course or

discipline. Another way is the EAREL scale (On-

line Learning Self-Regulation Scale), which focuses

on SRL strategies for distance learning activities, to

measure students’ self-regulation skills (Pierrot et al.,

2021).

3 METHODOLOGY

In this article, we carry out a secondary study to

identify, analyze, and interpret information related to

the research questions. We used the following ac-

tivities: planning, conducting, and reporting results

(Keele et al., 2007). First, a protocol for the review

was designed, which involves deﬁning the research

questions and ﬁltering the results based on previously

deﬁned criteria, in addition to removing duplicate ar-

ticles and the search term used. To assist in this pro-

cess, the tool parsing.al

was used to record the steps

used.

Search Terms: Based on the requirements, the

search term was proposed with the help of the Pop-

ulation, Intervention, Comparison, Result, and Con-

text (PICOC) structure. PICOC is used to formulate

research questions in systematic searches, as it covers

all the elements necessary to construct questions fo-

cusing on the real objective (Babar and Zhang, 2009).

We present in Table 2 the PICOC for this Systematic

Literature Review.

In total, three keywords in the English language

were used as search terms. First, keywords re-

lated to self-regulation of learning (”self-regulation”

OR ”self-regulated”). Afterward, keywords related

to online education and virtual learning environ-

ments (”e-learning” OR ”online education” OR ”on-

line learning” OR ”ITS” OR ”MOOC” OR ”LMS”).

And ﬁnally, the keywords related to recommendation

https://parsif.al/

Table 2: PICOC from Systematic Literature Review.

Aspect Value

Population (P) SR of learning, SR learning strategies

Intervention (I) Recommender Systems

Comparison (C) Other Literature Reviews

Results (O) Research where SRL is used to improve the

academic performance of students in RS.

Context (C) Online education, virtual learning environ-

ments, last 5 years.

systems (”recommendation systems”). The Search

String used has thus the following logical syntax:

(“self-regulation” OR “self-regulated”) AND (“e-

learning” OR “online education” OR “online learn-

ing” OR “ITS” OR “MOOC” OR “LMS”) AND

(“recommendation systems”).

Selection of Articles: The article selection pro-

cess involved the phases of the PRISMA Statement

(Moher et al., 2010) when instantiated to our review,

as shown in ﬁgure 1. Using the Search String we ob-

tained 272 results in the databases as given in Table 3.

Figure 1: Selection ﬂowchart used in this research based on

PRISMA (Moher et al., 2010).

Table 3: Number of Articles Selected by Search String from

the listed Databases.

List of Bases Number of Articles

ACM Digital Library 17

IEEE Digital Library 16

ISI Web of Science 53

ScienceDirect 56

Scopus 8

SpringerLink 122

Total 272

Among the 272 selected articles, 13 systematic lit-

erature reviews were found. Although several reviews

have been carried out to understand the ﬁeld of per-

The Use of Self-Regulation of Learning in Recommender Systems: State-of-the-Art and Research Opportunities

527

sonalized learning in recommender systems, only one

of them has focused on SRL, exploring its application

(Rasheed et al., 2020). In (Rasheed et al., 2020), a

systematic review is carried out to identify the chal-

lenges in the online component of hybrid teaching

from the perspective of students, teachers, and educa-

tional institutions. The main challenges students face

are related to self-regulation and the use of learning

technology.

In the article selection phase, we excluded 9 dupli-

cate articles, 1 article from the IEEE, 2 articles from

Web of Science, 5 articles from Scopus, and 1 ar-

ticle from SpringerLink; no articles from the ACM

and ScienceDirect were excluded at this stage. Selec-

tion criteria were developed to select articles that dis-

cuss SRL approaches and strategies in recommender

systems in the context of VLEs. According to the

research objectives, inclusion and exclusion criteria

were adopted, as shown in Table 4.

Based on the inclusion and exclusion criteria, 104

articles were selected by reading the title and sum-

mary, then we selected 42 articles based on reading

the introduction, methodology, results, and conclu-

sion. We ﬁnalized the selection of articles using the

inclusion and exclusion criteria and selected 8 arti-

cles, 3 articles from IEEE, 1 article from Web of Sci-

ence, 1 article from ScienceDirect, 1 article from Sco-

pus, and 2 articles from SpringerLink. Articles from

the ACM Digital Library were not selected, as they

were excluded based on the exclusion criteria.

To respond to the RQs, qualitative research was

carried out, which made it possible to obtain descrip-

tive data about students’ behavior regarding their way

of learning. Qualitative research is important in the

educational area, as it is essential to understand hu-

man reality, the difﬁculties experienced and the at-

titudes and behaviors of the subjects involved, thus

constituting essential theoretical support for educa-

tional research (Ferreira, 2015). In the data extrac-

tion phase, we extract data related to the research con-

text to infer whether a tool was used or proposed and

which tool was used or proposed; use of SRL and sys-

tem dynamics; the impact of students’ self-regulation

and interaction dynamics on VLE recommendations;

the learning self-regulation strategies carried out by

VLE students; how SRL was measured and what data

sets were used.

4 RESULTS STATE-OF-THE-ART

(RQ1-5)

The articles were selected based on the inclusion and

exclusion criteria, as well as the quality criteria. We

selected 8 articles, 3 articles from the IEEE Digital

Library, 1 article from the ISI Web of Science, 1 arti-

cle from ScienceDirect, 1 article from Scopus, and 2

articles from Springer Link, no article from the ACM

Digital Library was selected, as it was excluded based

on the exclusion criteria.

The articles selected in this review are presented

in Table 5. Taken together, these studies offer a snap-

shot of the state-of-the-art of the topic of interest here

and indicate a positive trend in the use of technolo-

gies to promote self-regulation by personalizing the

learning experience and providing valuable feedback

to students. However, it is important to recognize that

implementations must consider the diversity of educa-

tional contexts and the needs of individual students.

The articles included in this literature review ad-

dress the impacts and strategies of SRL on students’

academic performance, as well as how self-regulation

is measured, and which databases are used. In the

following subsections, we provide answers to the re-

search questions 1 to 5.

4.1 RQ1: What Is the Impact of

Student Self-Regulation and

Interaction Dynamics on

Recommendation in VLEs?

The use of SRL strategies has a signiﬁcant positive

impact on students’ interaction with the VLE, being

one of the main factors for the recommendations re-

ceived to be as assertive as possible, helping students’

academic performance. In (Odilinye and Popowich,

2020) study, student-generated metacognitive strate-

gies, such as highlighting and marking text, were nec-

essary to build a learning model that enabled appro-

priate personalized recommendations for completing

educational tasks. In (Wang et al., 2022), after adapt-

ing the existing VLE using the Personalized Quiz Al-

gorithm (PQ) and Knowledge Recommendation Al-

gorithm (KR), adaptively personalized quizzes and

recommendations were generated for individual stu-

dents, supporting students’ SRL.

Studies such as those by (Afzaal et al., 2021; Hi-

dayah et al., 2018; Wang et al., 2021) demonstrated

that SRL had a positive impact on the student’s aca-

demic performance in the courses. (Afzaal et al.,

2021) contributed by offering automatic, intelligent

recommendations developed from algorithms to help

students and teachers understand which resources a

student should work on to achieve the desired level

of performance. (Hidayah et al., 2018) contributed to

the generation of objective/sub-objective recommen-

dations, with recommendations for the use of strate-

CSEDU 2024 - 16th International Conference on Computer Supported Education

528

Table 4: Inclusion and Exclusion Criteria.

Inclusion Criteria Exclusion Criteria

1. Studies that present some approaches to self-regulation in learning environ-

ments.

2. Studies that provide empirical evidence on the advantages of using self-

regulation techniques.

3. Studies that focus on using SRL techniques to improve student’s learning ex-

perience and help teachers and tutors manage their students and groups.

4. Peer-reviewed studies that provide answers to research questions.

1. Duplicate studies.

2. Articles published before 2018.

3. Publications not related to the educational ﬁeld.

4. Studies that are not related to VLE.

5. Studies not related to recommendation systems.

6. Studies that do not present approaches to SRL.

7. Non-peer-reviewed studies.

8. Secondary studies.

9. Gray literature.

Table 5: Selected Articles.

Article Authors Year

Automatic and Intelligent Recommendations to Support Students’ Self-Regulation Afzaal, Nouri, Zia, Papapetrou, Fors, Wu, Li, and Wee-

gar.

2021

Facilitating English Grammar Learning by a Personalized Mobile-Assisted System

with a Self-Regulated Learning Mechanism

Wang, Chen and Zhang 2021

Personalized Recommender System Using Learners’ Metacognitive Reading Ac-

tivities

Odilinye and Popowich 2020

Process mining for self-regulated learning assessment in e-learning Cerezo, Bogar

ın, Esteban, and Romero. 2020

Promoting self-regulated learning strategies for ﬁrst-year students through the

COMPER service

Pierrot, Michel, Broisin, Guin, Lefevre, and Venant 2021

The relationship between self-regulated student use of a virtual learning environ-

ment for algebra and student achievement: An examination of the role of teacher

orchestration

Leite, Kuang, Jing, Xing, Cavanaugh, and Huggins-

Manley

2022

A Framework for Improving Recommendation in Adaptive Metacognitive Scaf-

folding

Hidayah, Adji and Setiawan 2018

IFSE - Personalized Quiz Generator and Intelligent Knowledge Recommendation Wang, Li, Zimmermann, Pinkwart, Werde, Van Rijn,

DeWitt, and Baudach

2022

gies that contribute to students’ SRL and the improve-

ment of academic performance.

In (Wang et al., 2021), results showed that par-

ticipants who used SRL strategies signiﬁcantly out-

performed the participants of the control group in

test scores. In (Cerezo et al., 2020), the develop-

ment of alert systems to predict students at risk dur-

ing a course, and the personalization of self-regulated

VLEs, with the construction of RS based on differ-

ent SRL behaviors, have a positive impact on student

academic performance. Consistent evidence from

these studies reinforces the conclusion that promoting

SRL is important to optimize the learning experience

in VLEs, providing more effective recommendations

and directly contributing to students’ academic suc-

cess.

4.2 RQ2: Which SRL Strategies Are

Carried Out by Students in a VLE?

As the answer to RQ1 indicates, the use of SRL strate-

gies promotes an increase in student academic per-

formance. Students use SRL strategies to evaluate

their experience in the course, motivation to com-

plete the course, carry out tasks, analyze the time to

complete the task, and analyze the completion and

grades obtained in the course (Afzaal et al., 2021).

They also carry out strategies related to deﬁning their

learning goal, deciding the level of the learning ma-

terial, choosing between reviewing previously incor-

rectly answered questions or new questions, and re-

ceiving a report with their performance for reﬂection

on learning (Wang et al., 2021).

Students carry out metacognitive reading activ-

ity strategies (text marking, tags) to extract the most

relevant information from the text (Odilinye and

Popowich, 2020), action strategies indicative of un-

derstanding and learning the materials, implementa-

tion and review actions (Cerezo et al., 2020), strate-

gies for organizing the learning context and request-

ing peer support (Pierrot et al., 2021), strategies

for monitoring performance and self-adaptation tasks

(Wang et al., 2022). In (Leite et al., 2022) and

(Hidayah et al., 2018) they used metacognitive self-

regulation strategies such as monitoring, effort regu-

lation, and metacognition as self-testing (Leite et al.,

2022).

The use of SRL strategies plays a crucial role in

increasing students’ academic performance in VLEs.

The Use of Self-Regulation of Learning in Recommender Systems: State-of-the-Art and Research Opportunities

529

The diversity and scope of these strategies high-

light the importance of promoting educational envi-

ronments that not only recognize, but also actively

encourage self-regulation, empowering students to

shape their own learning experience and achieve more

meaningful academic outcomes.

4.3 RQ3: How Does a Student’S

Current Posture Inﬂuence Their

Ability to Self-Regulate Their

Learning in a VLE?

Research presents some attitudes of students that in-

ﬂuence their ability to self-regulate. According to

(Pierrot et al., 2021), dropout students do not use any

SRL strategy, they procrastinate and communicate lit-

tle with their peers, while follower students use some

strategies, but procrastinate and start working by com-

municating with their peers. However, solitary per-

formers use strategies, do not procrastinate, and do

not communicate with their peers. Finally, effective

students use self-regulatory strategies, do not procras-

tinate, and communicate with their peers. By consid-

ering students’ diverse stances toward self-regulation,

educators can develop more targeted and personalized

strategies to effectively support students’ academic

development and self-regulation in educational set-

tings.

4.4 RQ4: How to Measure SRL?

SRL can be measured by analyzing student data in

the VLE. In (Afzaal et al., 2021), SRL was evalu-

ated by analyzing student performance in a program-

ming course. Initially, students’ experience and mo-

tivation were assessed, followed by an analysis of

the attributes of the questionnaire and tasks related

to scoring and time spent, then the attributes of ac-

tivity completion were examined, including count of

video views, materials, and forums, followed by an

analysis of student grades. To predict future perfor-

mance, tests were carried out with Artiﬁcial Intelli-

gence algorithms, with the Artiﬁcial Neural Network

(ANN) outperforming the others in all measures, al-

though Random Forest (RF) was similar in predicting

questionnaires, and K-Nearest Neighbors (KNN) and

Support Vector Machines (SVM) produced identical

results across all tasks. On the other hand, Logistic

Regression (LR) performed worse than the other al-

gorithms. (Hidayah et al., 2018) also measured SRL

by analyzing student data in the VLE. From the stu-

dent’s interaction with the system and the deﬁnition

of their objectives, it was possible to develop the stu-

dents’ modeling in their system.

Another way to measure SRL, which also uses the

analysis of student data in the VLE, is by extracting

student records. In the (Cerezo et al., 2020) study,

student records were extracted and related to four at-

tributes: time, student identiﬁers (ID), action, and in-

formation. SRL can also be measured by collecting

types of learning information from students. In (Wang

et al., 2022), it was measured using the PQ and KR al-

gorithm; these algorithms separate questions and quiz

options, making it easy to automatically generate per-

sonalized quizzes for each student. Student perfor-

mance is monitored through responses to quizzes; the

separation between questions and options allows for

the dynamic creation of custom quizzes and ques-

tions. The system provides adaptive feedback based

on the student’s knowledge by connecting knowledge

concepts directly to quiz options. Learning materi-

als are linked to quiz options, making it easy to iden-

tify knowledge gaps or errors. The system also pro-

vides accurate feedback for wrong answers and rec-

ommends additional content when students respond

correctly.

Data analysis in VLE emerges as a versatile and

effective tool for measuring SRL. The combination

of objective methods, such as Machine Learning al-

gorithms, with subjective approaches, such as ques-

tionnaires and scales, provides a comprehensive and

meaningful view of the students’ self-regulation pro-

cess in the VLE. The results can guide more personal-

ized and effective pedagogical practices and teaching

strategies.

4.5 RQ5: Which Datasets Are Used in

Research?

All articles used real student data collected from

VLEs with information about students’ educational

activities during a course. The VLEs used were Moo-

dle (Cerezo et al., 2020; Wang et al., 2022), nStudy

(Odilinye and Popowich, 2020), exercise platform

(Pierrot et al., 2021), personalized assisted mobile

system (Wang et al., 2021) and Math Nation (Leite

et al., 2022). (Afzaal et al., 2021) and (Hidayah et al.,

2018), did not specify the name of the VLE used.

In (Afzaal et al., 2021; Hidayah et al., 2018; Pier-

rot et al., 2021) VLE data from students in the com-

puting area were used. In (Wang et al., 2021) data

from pre-test and post-test scores were used, from

randomly selected students. (Cerezo et al., 2020)’s

research used data from undergraduate students from

an online course on Moodle. VLE logs were ex-

tracted from real events (time, student ID (to maintain

anonymity), action, and information), which were rel-

CSEDU 2024 - 16th International Conference on Computer Supported Education

530

evant to the process of self-regulation of learning and

academic performance of the course.

The (Hidayah et al., 2018) research also used stu-

dent interaction log data in the VLE, but the VLE

used was another unspeciﬁed one. In (Leite et al.,

2022) data from Math Nation was utilized and inte-

grated into the student information system. In (Wang

et al., 2022) two datasets were used, a small dataset

with 1,000 students and 10,000 question options, and

a large dataset with 10,000 students and 100,000 op-

tions. In (Odilinye and Popowich, 2020) data from

49 undergraduate students from a Canadian university

were used.

Data collection in VLEs provides a solid basis

for investigating SRL in diverse educational environ-

ments. The interdisciplinary approach and the variety

of analyzed data contribute to a comprehensive under-

standing of SRL, providing valuable information for

the continuous improvement of teaching and learning

methods.

5 RESULTS - RESEARCH

OPPORTUNITIES (RQ6)

In this section, we check what the selected/reviewed

papers suggest as future work and identify the oppor-

tunities found for future research.

One suggestion is to use experiments on larger

data sets and collaboration with teachers to determine

the effectiveness of the proposals presented (Afzaal

et al., 2021). Another is to implement and test the

functionality of recommending the use of SRL strate-

gies in a classroom environment with students (Hi-

dayah et al., 2018). More research is needed to under-

stand which design features lead students to believe

which visualization is easier to use. Better under-

stand students’ motivations for using these services

and how best to adapt design and implementation to

their needs (Pierrot et al., 2021). Future research can

investigate how the integration of other VLE func-

tionalities can be included in a personalized learning

recommendation system, such as collaborative learn-

ing and question generation module (Odilinye and

Popowich, 2020).

One need is to incorporate more natural language

processing functions into the VLE. Teachers can be

assisted with questions about a given domain. Collect

new types of student learning information for deeper

machine learning analysis, such as the time taken to

answer questions, feedback, and number of hits from

recommended resource links (Wang et al., 2022). An-

other future work is to shift focus to other relevant

VLEs, such as MOOCs, and check ﬁndings across

different types of learning platforms (Cerezo et al.,

2020).

It is also important for future studies to consider

students’ learning characteristics, such as their cog-

nitive learning styles or types of SRL and examine

whether students of various learning proﬁles would

beneﬁt differently from this system (Wang et al.,

2021). Other SRL strategies, such as help-seeking

or peer learning, can be used. Future research could

include more SRL strategies and investigate whether

teacher instrumental orchestration continues to mod-

erate the relationship between student SRL and stu-

dent achievement (Leite et al., 2022).

We detected that self-regulation construct surveys

have been carried out using several questionnaires, of-

ten adapted to the speciﬁc needs of the authors. A

speciﬁc SRL questionnaire for VLEs that can brieﬂy

calculate these indicators could be a point of advance-

ment in research that uses SRL. Another possibility is

to integrate VLEs with SR mechanisms and recom-

mend educational objects to improve student perfor-

mance.

6 FINAL CONSIDERATIONS

The application of metacognitive strategies and the

development of adaptive algorithms result in person-

alized recommendations, positively inﬂuencing aca-

demic performance and strengthening SRL. SRL not

only affects test scores, but also contributes to the

achievement of speciﬁc educational goals.

Projects such as alert systems and personalization

of online environments based on SRL behaviors have

a positive impact by predicting at-risk students and

providing a more effective learning experience. Mea-

suring learning Self-regulated is possible through the

analysis of data generated in VLEs. Studies highlight

the continuous need to integrate and improve SRL

strategies in VLEs. Students’ active and conscious

promotion of SRL contributes not only to academic

development but also reﬂects an engaged, in-depth

engagement with the learning material.

This article provides a comprehensive overview

of SRL use over the last 5+ years (2018-2023),

highlighting measures and strategies for assessment.

Its ﬁndings stemmed from answers to 6 Research

Questions. Taken together, the answers provided an

overview of the state of the art and supported an in-

dication of research opportunities. The main ﬁndings

and opportunities for research were as follows. Un-

derstanding students’ motivations for using VLEs and

personalizing the design and implementation accord-

ing to their needs; The incorporation of NLP func-

The Use of Self-Regulation of Learning in Recommender Systems: State-of-the-Art and Research Opportunities

531

tions in VLEs Future research should focus on how

these NLP can support teachers in this process.

This work contributed to providing more insight

into how SRL has been used over the years in educa-

tion, seeking to highlight how self-regulation is mea-

sured, which self-regulation strategies are used and

what is the impact of SRL on student performance.

Regarding future work, we hope to see more experi-

ments on improving student performance and motiva-

tion using SRL strategies.

REFERENCES

Afzaal, M., Nouri, J., Zia, A., Papapetrou, P., Fors, U., Wu,

Y., Li, X., and Weegar, R. (2021). Automatic and in-

telligent recommendations to support students’ self-

regulation. In 2021 International Conference on Ad-

vanced Learning Technologies (ICALT), pages 336–

338. IEEE.

Al-Obaydi, L. H. (2020). Using virtual learning environ-

ment as a medium of instruction in eﬂ context: Col-

lege teachers’ attitudes. Intensive Journal, 3(2):18–

30.

Babar, M. A. and Zhang, H. (2009). Systematic literature

reviews in software engineering: Preliminary results

from interviews with researchers. In 2009 3rd Interna-

tional Symposium on Empirical Software Engineering

and Measurement, pages 346–355. IEEE.

Ben-Eliyahu, A. and Linnenbrink-Garcia, L. (2015). In-

tegrating the regulation of affect, behavior, and cog-

nition into self-regulated learning paradigms among

secondary and post-secondary students. Metacogni-

tion and learning, 10:15–42.

Boruchovitch, E. (2014). Autorregulac¸

ao da aprendiza-

gem: contribuic¸

oes da psicologia educacional para a

formac¸

ao de professores. Psicologia Escolar e Edu-

cacional, 18:401–409.

Brito, P. H., Bittencourt, I. I., Machado, A. P., Costa, E.,

Holanda, O., Ferreira, R., and Ribeiro, T. (2014). A

systematic approach for designing educational recom-

mender systems. In Software Design and Develop-

ment: Concepts, Methodologies, Tools, and Applica-

tions, pages 1264–1288. IGI Global.

Cerezo, R., Bogar

ın, A., Esteban, M., and Romero, C.

(2020). Process mining for self-regulated learning

assessment in e-learning. Journal of Computing in

Higher Education, 32(1):74–88.

Dwivedi, S. and Roshni, V. K. (2017). Recommender sys-

tem for big data in education. In 2017 5th National

Conference on E-Learning & E-Learning Technolo-

gies (ELELTECH), pages 1–4. IEEE.

Ferreira, C. A. L. (2015). Pesquisa quantitativa e qualita-

tiva: perspectivas para o campo da educac¸

ao. Revista

Mosaico-Revista de Hist

oria, 8(2):113–121.

Hidayah, I., Adji, T., and Setiawan, N. (2018). A framework

for improving recommendation in adaptive metacog-

nitive scaffolding. In 2018 4th International Confer-

ence on Science and Technology (ICST), pages 1–5.

IEEE.

Keele, S. et al. (2007). Guidelines for performing system-

atic literature reviews in software engineering.

Kuo, Y.-C., Walker, A. E., Schroder, K. E., and Belland,

B. R. (2014). Interaction, internet self-efﬁcacy, and

self-regulated learning as predictors of student satis-

faction in online education courses. The internet and

higher education, 20:35–50.

Leite, W. L., Kuang, H., Jing, Z., Xing, W., Cavanaugh, C.,

and Huggins-Manley, A. C. (2022). The relationship

between self-regulated student use of a virtual learn-

ing environment for algebra and student achievement:

An examination of the role of teacher orchestration.

Computers & Education, 191:104615.

Lima, J. C. M. P. d. and Silva, B. D. d. (2010). Tic e pro-

cessos de auto-regulac¸

ao da aprendizagem.

McLellan, C. K. and Jackson, D. L. (2017). Personality,

self-regulated learning, and academic entitlement. So-

cial Psychology of Education, 20:159–178.

Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., Group,

P., et al. (2010). Preferred reporting items for system-

atic reviews and meta-analyses: the prisma statement.

International journal of surgery, 8(5):336–341.

Obeid, C., Lahoud, I., El Khoury, H., and Champin, P.-

A. (2018). Ontology-based recommender system in

higher education. In Companion proceedings of the

the web conference 2018, pages 1031–1034.

Odilinye, L. and Popowich, F. (2020). Personalized rec-

ommender system using learners’ metacognitive read-

ing activities. In Methodologies and Intelligent Sys-

tems for Technology Enhanced Learning, 10th Inter-

national Conference, pages 195–205. Springer.

Pierrot, L., Michel, C., Broisin, J., Guin, N., Lefevre,

M., and Venant, R. (2021). Promoting self-regulated

learning strategies for ﬁrst-year students through the

comper service. International Association for Devel-

opment of the Information Society.

Pintrich, P. R. et al. (1991). A manual for the use of the

motivated strategies for learning questionnaire (mslq).

Polydoro, S. A. J. and Azzi, R. G. (2009). Autorregulac¸

da aprendizagem na perspectiva da teoria sociocog-

nitiva: introduzindo modelos de investigac¸

ao e

intervenc¸

ao. Psicologia da Educac¸

ao, (29).

Rasheed, R. A., Kamsin, A., and Abdullah, N. A. (2020).

Challenges in the online component of blended learn-

ing: A systematic review. Computers & Education,

144:103701.

Soares, L. T. (2018). Aprendizagem autorregulada: uma

revis

ao bibliogr

aﬁca. Anais do SIMPOM, (5).

Takami, K., Dai, Y., Flanagan, B., and Ogata, H. (2022).

Educational explainable recommender usage and its

effectiveness in high school summer vacation assign-

ment. In LAK22: 12th International Learning Analyt-

ics and Knowledge Conference, pages 458–464.

Wang, X., Chen, J., and Zhang, T. (2021). Facilitating

english grammar learning by a personalized mobile-

assisted system with a self-regulated learning mecha-

nism. Frontiers in Psychology, 12:624430.

Wang, X., Li, H., Zimmermann, A., Pinkwart, N., Werde,

S., Van Rijn, L., De Witt, C., and Baudach, B.

(2022). Ifse-personalized quiz generator and intel-

ligent knowledge recommendation. In 2022 IEEE

16th International Conference on Semantic Comput-

ing (ICSC), pages 201–208. IEEE.

CSEDU 2024 - 16th International Conference on Computer Supported Education

532