An approach to providing recommendations for the selection of learning resources for various learning purposes is proposed, taking into account preferences for learning styles, such as audio, text, etc. A framework for constructing a recommendation system is presented, which enables generating personalized recommendations for students, considering their accumulated learning experience. Incorporating such recommendations when choosing resources can make learning more comfortable, as it allows the selection of materials aligned with individual preferences in presentation.
The task of preparing learning recommendations is aimed at improving the process of acquiring knowledge and skills. It can help learners select resources that match both their learning style and their goals.
Therefore, the development of a recommendation system capable of automatically generating suggestions for choosing learning resources, while considering the preferred style of content presentation, is an urgent and relevant task.
The purpose of this paper is to propose a methodology for generating recommendations for selecting learning resources based on learning styles, with the aim of improving the overall comfort and effectiveness of learning.
The paper is structured as follows. The first section describes the task of digital learning and the problems that arise there. The second section describes the task of preparing recommendations for selecting educational resources. The third section outlines the methodology for generating such recommendations.
The movement for open educational resources [
The emergence of new technologies has led to the need for continuous updating of knowledge,
acquisition of new skills, and improvement of qualifications [
Modern technologies allow the creation of multimedia and interactive content even by
individuals without extensive training in this field. For example, teachers use YouTube to deliver
lessons, while H5P enables the quick development of tests and other learning content [
The application of recommender systems in education requires considering a wide range of approaches in recommender systems rely on user-generated data, such as ratings, reviews, and feedback.
In [
Study [
The authors of [
So, today, when a large number of various courses for learning in various fields have been developed, which can be used both for independent learning and for specialized training, it is very difficult for a student to navigate in choosing an educational resource, even taking into account the purpose of his learning. Therefore, building technologies that could help him in this choice is an important and urgent task today. This study aims to develop a methodology for preparing recommendations for a student, taking into account a comfortable learning style.
Let's take a closer look at what educational resources exist today.
Educational resources can be presented in digital (electronic) and traditional formats, including electronic textbooks, interactive assignments, educational videos, and audio, presentations, educational and supportive software, including games and as well as books, reference books, and other materials used in the educational process.
The creation of standardized content for use in a variety of learning environments: face-to-face, blended, and/or virtual, is critical in today's education.
The learning object approach aims to facilitate the development of small units of content that can be combined and reused across courses, thereby taking advantage of the development of educational programs and materials.
According to the IEEE Educational Technology Standards Committee [
A handy tool for creating learning resources is H5P [
Educational resources are a set of materials, tools, platforms, and services aimed at supporting and optimizing the learning process. This concept covers both traditional educational and methodological complexes and innovative digital solutions: from electronic textbooks to adaptive learning systems and virtual reality. The following gives a summary of all heading levels.
Choosing the optimal educational resource is a task that requires a systematic approach. With a huge variety of available platforms, it is easy to get lost or make a choice that does not correspond to real needs. The key to success is a clear formulation of educational goals and a critical analysis of the options offered.
To select the appropriate educational resources from a large volume of offered ones, a recommendation that could guide the student in choosing educational resources would be very useful for the student.
Such a recommendation could be obtained using a recommendation system that, based on the collected information about the courses and the students who have completed them, would prepare recommendations for new students, taking into account the situation in which the student selects a resource for learning. This would allow the student's time to be used more efficiently in choosing a resource and selecting a resource that best suits the student's needs, which would improve the quality of the acquired skills or knowledge.
Such a recommendation system can be considered as an advisor that supports decision-making
based on information about the student and their experience in a particular field of knowledge.
Most educational recommendation systems use available information, such as descriptions of
educational resources, student characteristics (age, gender, group, learning style, grades, etc.).
Recommendations from such systems often assume similarities between students within a group,
or similarities in the discipline being studied [
The quality of the recommendations that the recommendation system produces may depend on the information available about the students and their resources for learning. The more complete the information is, the more likely it is that the result will be most suitable for the student.
The situations in which a person who chooses a course for learning may find himself may be different. The authors conducted a study that showed that four types of situations can be distinguished in lifelong learning. For each of the situations, different methods and recommendations can be applied. Let us consider such situations in more detail.
For example, let's consider situations in which the need to select an educational resource arises.
Situations in which students may find themselves when choosing a learning resource for a lifelong
learning case were investigated [
The authors identified situations based on achieving the following goals: for professional development: the learner needs a proven learning outcome in a certain area, for example, in programming. In this case, the resource represents a course or program, activities, assignments, and assessments leading to confirmation of the learner's competence (diploma or certificate); or personal development: the person needs to learn how to perform a task or solve a problem.
Resource recommendations may include various types of resources for acquiring knowledge or skills, including microlearning videos; satisfying one's own curiosity: this may be simply a hobby, a need to clarify details, improve understanding of something, or simply a desire to keep up with modern technological advances. Various resources can be recommended here, from educational and entertaining to scientific research, passive presentations, interactive videos, games, and quizzes; to consolidate existing skills or refresh knowledge: this is a specific situation that has not yet been sufficiently studied. This is explained by the fact that, regardless of the subject area, acquired knowledge degrades over time and requires updating. It may also be necessary to work on the acquired skills.
You can also divide the learning process into superficial, when it is enough to repeat an exercise once or twice, or in-depth, when you need to return to repeating individual fragments or repeating exercises.
In addition, each student has preferences for the style of presentation of learning materials. We will consider this in more detail in the next section.
Learning style is usually understood as the preferred way of accepting and processing new information.
There are several approaches to describing learning styles. Learning style is understood as a set of parameters that characterize the method of presenting and perceiving information, the method of awareness, and cognitive characteristics. Using a preferred style promotes effective assimilation. There are several models that systematize the characteristics of the learner that determine his or her learning style.
For example, such as Learning Styles VARK [
Other learning styles are also known, such as Kolb learning [
The Felder-Silverman (FS) model describes learning preferences as a set of parameters across four dimensions that characterize the best way the new information is both perceived and processed. The drawback of the classical FS model comparing to VARK is in separation between verbal and visual information without further detailization to text or audio presentation. information processing (sequential vs global).
The learning styles FS model describes four measurement scales that determine students' learning preferences: active/reflective, sensory/intuitive, visual/verbal, and sequential/global.
This model helps teachers adapt teaching methods to different student styles, making the learning process more comfortable.
FS model scales: 1 scale. The learning process can be active or reflective, i.e. it is an action, not a reflection.
Active learners prefer to learn by actively participating in the process, for example, discussing, working in groups, conducting experiments. Reflective learners prefer to think about information, analyze it and reflect on it.
2 scale. Perception in the learning process is divided into sensory or intuitive, which can be understood as concrete and conceptual, as fact and theory.
Sensory learners prefer facts and concrete details, while intuitive learners prefer abstract concepts and connections between them.
3 scale. Information input is divided into visual or verbal. Images and diagrams are visual information, reading/writing is verbal.
Visual learners learn information better when presented in the form of images, diagrams and charts, while verbal learners prefer written and oral explanations.
4 scale. Understanding information can be sequential or global. That is, either a step-by-step, orderly understanding of the material or broad thinking.
Sequential learners prefer a logical and linear approach to learning, assimilating information step by step, while global learners prefer to see the big picture and then understand the details.
Each area can be understood as a separate approach to learning. learning content and arranging the learning process.
In group learning and in creation of online courses the diversity of learning styles is usually addressed though it may lead to unnecessary duplication of information.
1. To adapt learning materials:
Teachers can use the model to develop learning materials that suit different learning styles. For example, more interactive activities can be added for active learners, and more hands-on examples can be added for sensory learners.
2. Diversification of teaching methods:
The model helps to vary the teaching methods to suit the needs of different learners. For example, lectures, group discussions, practical work, and independent study can be used. 3. Increased student engagement:
Recognizing and accommodating different learning styles can increase student engagement in the learning process, as they will feel that their needs are being taken into account.
The FS model provides a valuable tool for understanding and satisfying different learning styles, which contributes to a more comfortable and enjoyable learning process.
It is assumed that the learning style will not change throughout the learning content. However, informal discussion in the focus group revealed significant changes in the parameters depending on the learning situation.
The combination of the parameters (styles) of the FS model forms individual learning preferences.
Now it is turn to the methodology for obtaining recommendations using the learning style.
Recently, personalization of information has become a key factor, as it allows the selection of information taking into account the interests of the user.
Continuous electronic learning is becoming increasingly popular, as it allows people of different ages to study in different conditions, both to obtain a diploma and simply to refresh their knowledge in some subjects. Taking into account the personal preferences of the student can make learning more comfortable, and there is a greater likelihood of completing the course.
To select suitable educational resources from a large volume of offered ones, a recommendation system would be very useful, which, based on the collected information about the courses and students who have completed them, would prepare recommendations for new students, taking into account the situation in which the student selects a resource for training. This would allow more efficient use of the student's time to select a resource and select a resource that best suits the student's needs, which would improve the quality of the skills or knowledge acquired.
The quality of recommendations issued by a recommendation system may depend on the available information about students and learning resources. The more complete the information is, the more likely it is that the result will be most suitable for the student.
The task of preparing recommendations during training is aimed at improving the process of acquiring knowledge and skills. This can help the student select resources for training, considering the style and purpose of training. Therefore, the development of a recommendation system that would allow automatic preparation of recommendations for choosing a training resource for a specific user, taking into account the style of presentation of materials that is comfortable for him, is a relevant task.
In this study, we consider resources where the author of each of provides their characteristics.
The metadata structure recommended by the standard [
Thus, the user receives a recommendation, the reason for which can be known and explained. It is related to a specific educational situation and, therefore, will help the user get what he would like with a greater probability.
This approach can be useful for solving the cold start problem. First, the resources already have an initial description that is reliable and not specific to the user. It is assumed that users provide information about their preferences specific to a particular situation, but this information (template) is independent of the resource. If a complete template is missing, the most important parameters can be requested separately and processed individually.
We will assume that the preliminary selection of resources has already been completed and the resources analyzed for compliance with the style have matched it (meaning the resource topic and language, and possibly the duration of study, price and other critical characteristics).
To generate recommendations, information is used both on the properties of the resource and on the preferences of users. Since the discussion of preferences within our small group showed that the characteristics of the style depend on the situation - the purpose of study, then we will link the corresponding user profile to the situation and request confirmation the next time the resource is requested for a recommendation.
We propose using individual learning style preferences specific to a particular situation. The methodology for preparing recommendations involves stages, the sequence of which is shown in Fig. 1.
Let's consider a small example of how recommendations differ depending on preferences.
Suppose we have information about 10 courses presented by a certain university.
Suppose that users who have completed some courses have chosen the following styles recommendations: • • •
Kate - Global, Visual, Reflective; Olha - Active, Visual, Sequential.
Comments about the courses of users who have already completed them can also be presented. These comments can be recommendations for new users who are choosing a course for themselves to study.
For example, Comments Kate: I liked the recommendation of the resource for repetition, because it briefly and clearly shows the use of prepositions.
A new user who forms his preferences about the learning style will be closer to one of the presented users who have already completed these courses. Then the recommendation system will prepare recommendations for the new student about three courses, recommended by one of the recommendations that is closer to the student in terms of learning style.
This technique will be discussed in more detail with examples in subsequent articles. 10. Discussion
Although group learning, including distance learning, seeks to address the needs of learners with different styles whenever possible, it is assumed that the use of resources aligned with an process an aspect particularly important in lifelong learning.
The choice of the FS model for describing learning styles is justified by the fact that it accounts not only for the form of information presentation but also for the ways in which learners engage with it.
During the study, the need to adapt the FS model was identified, specifically: • • clarifying the form of verbal information presentation (text or audio); considering the learning objective (situation) when searching for educational resources.
As the discussion indicates, both the choice of information presentation format and the
preferred learning activities depend on the purpose of accessing a resource. The diversity of
learning situations explains the use of scales to determine the values of individual style parameters.
For example, the standard preference may be verbal (text), but for repetition, a combination of
visual and audio modalities is more effective. Such an interpretation provides richer insights than
relying solely on numerical values [
The methodology proposed by the authors allows, based on descriptions of a set of resources provided by developers, offering learners the three most suitable resources that match their preferred learning style.
This paper proposes an approach to providing recommendations for selecting educational resources for various learning objectives, taking into account students' learning style preferences. A framework for constructing a recommender system has been developed that enables the generation of personalized recommendations for students based on their accumulated learning experience. Incorporating such recommendations into resource selection can improve the learning experience by accommodating students' individual preferences.
Future research will focus on implementing the proposed recommendation system and testing it on specific use cases, providing tailored recommendations for various learning purposes.
During the preparation of this work, the authors have not employed any Generative AI tools.