The objective of this study is to perform a comparative analysis of fuzzy controllers and selected machine learning algorithms in adaptive e-learning systems for solving classification problems characterized by imprecisely modeled input data. In scenarios where access to labeled datasets for solving classification problems is not available, and thus the application of machine learning methods is not feasible, expert knowledge encoded in the form of if...then rules can serve as a basis for designing an expert fuzzy controller. The paper studies a classification problem where the expert fuzzy controller approach was applied and evaluates its performance against the most widely used machine learning algorithms. In addition, the study includes a comparison with fuzzy controllers generated automatically from an available dataset. The process of data representation for machine learning purposes is also outlined. The results show that machine learning algorithms achieve the highest classification accuracy when a dataset is available. However, fuzzy controllers demonstrate competitive performance, highlighting their potential utility in situations where datasets are not present.

Learning materials provided to students during remote learning are characterized by different technical forms and interactivity levels. The structure of the materials is often enforced by the tools used for their creation, it could also result from the preferences and experiences of their authors. This article presents the effect of the structure of learning materials on the student's performance as well as the dynamics of all interactions of students with such materials. The experiment used learning materials provided to students for self-study as part of classes from a 15-week database course. The classes were held in the form of weekly lectures and teaching labs. The course comprised several larger sections, beginning with the introduction to the SQL SELECT command. Each learning materials module for this part of the course contained approx. 25 programming tasks which were assessed automatically using the Jobe/CodeRunner environment. It ended with a quiz with random questions of single- or multiple-choice, fill-in the gap, or calculate the result. Before the course started, 88 students were randomly divided into two groups, each numbering 44 students. The first group received learning materials with a granular structure, which were divided into small, thematically cohesive chunks of knowledge. The second group received the same learning materials organized in the manner webpages are constructed (long form), i.e., all the material content from one class was presented on one webpage. This is a solution known, for example, from Jupyter Notebook documents. The deadline for completing one module and taking the quiz at the end of it was one week and was the same for both groups. The self-learning materials replaced traditional weekly teaching labs, but students were allowed to contact the teachers remotely to obtain help or additional explanations. The results show that the structure of the materials did not have a significant effect on the performance of the students, yet they were more involved in learning and spent more time in the course when they worked with granular materials. Their levels of satisfaction were also higher. The results presented in the article may be of interest to teachers when deciding on the form of learning materials to be created.

This article presents an e-learning system that supports the adaptation of course content to the student's way of learning. The adaptability of an e-learning course is a challenging task that requires appropriate tools and a well-thought-out architecture. The solution presented in this article is based on the concept of information imprecision and fuzzy control. The proposed fuzzy controller, based on an expert rule base calculates the student's activity in the course in order to determine the questions in tests at the end of successive parts of the course. The solution proposed was implemented in an e-learning Applications of Fuzzy Sets course, which was part of an Artificial Intelligence class. The proposed system successfully adapts the course content to the student's learning style. The article highlights the potential of fuzzy controllers in adaptive e-learning systems. The use of fuzzy controllers makes it possible to take into account the imprecision of information, enabling better handling of uncertainty and variability in the student's learning style.

The outbreak of the COVID-19 pandemic gave rise to a need to change course syllabi in order to completely transition to a remote learning model. In the case of subjects comprising programming tasks, taking into account the availability of tools and resources that provide opportunities for independent work without teacher supervision, it was necessary to decide whether students should be allowed to work under self-paced learning, or, similar to traditional classes, fixed-schedule learning. Experiences from MOOCs demonstrated that course scheduling is not without impact on the learning process, and may affect the level of student satisfaction with the course. This study determines how course scheduling affected the performance of students attending a database course. The students were divided into two groups, and completed the first module without teacher supervision. During the learning process, they solved programming tasks which were assessed automatically, and then took quizzes to verify what they learned. One of the groups worked with the materials and took the quizzes in accordance with a schedule, and the other group did so without any time constraints. The results demonstrate that students perform better when working under the fixed-schedule model, without any impact on their level of satisfaction with the course. The system of learning not only affected the quiz results in the module where different scheduling was used, but student performance in later parts of the course as well. The results presented in the paper should be of interest to teachers designing remote courses involving self-learning.

The article presents a solution supporting individualised learning in courses with a tight schedule. Such courses pose additional organisational challenges and require appropriate tools. The presented solution is based on an Intelligent Tutoring System immersed in repository of e-learning content, which enables selection of content immediately before its provision to the student instead of at the beginning of a course. Thanks to this, the system, having identified the student's needs, is able to make available the most suitable repository content at a given stage of education. The flexibility of the system is guaranteed by modularisation of content and its logical division using the UCTS taxonomy. The content has been described by means of concepts arranged according to the specificity of the domain to which the resources belong in order to ensure that the ITS is able to select relevant content. The proposed solution was used to set up an Applications of Fuzzy Logic course, which was part of an Artificial Intelligence class. The course was conducted within a very limited time frame resulting from the COVID-19 epidemic.

Abstract The article presents the principles for creating intelligent tutoring systems, which are active in dynamically growing repositories of e-learning contents. To make such a system work it is necessary to give the contents appropriate structure to enable it to be used in numerous educational contexts. The architecture of intelligent tutoring system is based on wordnet based ontology with expert knowledge, which has been used for repository resource indexing, and which is a basic component of domain model. The solution in its entirety has been supplemented with software agent, responsible for linking the most relevant content in repository to predefined educational strategy.

The paper presents the method and tools to build repositories of digitalized didactic materials to support their reusability. The method consists of a set of recommendations for the structuring of didactic materials and of a way to assign unambiguous didactic interpretation to sections of materials by means of UCTS. UCTS (i.e. Universal Curricular Taxonomy System) is a taxonomic system designed to situate learning content in the didactic process. Using this system to mark parts of materials ensures that users will access materials that are cohesive and relevant without the risk of downloading from the repository incomplete chunks of content. Such guarantee is indispensable in massive repositories, which are created and used by numerous users, and which contain non-uniform content. An important component of the presented method is dedicated software Content Repository Tool, which was designed to store and process didactic content with the ability to reuse it in different educational contexts. This software enables creation of single- or multi-topic repositories with the possibility to create distributed repositories.

This paper introduces the issues of cartographical presentation of spatial data in indoor navigation systems. The research covers ensuring appropriate map communication process from modeling spatial data (creating databases) to formulating a cartographical message for the final recipient. The use of construction drawings and architectural plans in the process of visualization is a considerable limitation and can be applied only in the testing phase of navigation systems. Indoor plans created by graphic designers also limit the functionality of the system. For this reason, it is necessary to apply GIS vector data models and cartographical methodology in the development process of maps dedicated to indoor navigation purposes. Moreover, what poses a vital issue is the presentation of a building at many levels of detail (data generalization) and appropriate symbolization designed for displaying on the small screen of a mobile device during the user's movement. In a spatial database, apart from a classic navigation graph, additional vector data of probable indoor movement traces should be saved, which would allow for generation of navigation directions as well as optimal cartographical visualization of calculated routes.