| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

Search the digital library catalog Help

Query: search in
search in
search in
search in
* old and bologna study programme

Options:
  Reset


1 - 10 / 193
First pagePrevious page12345678910Next pageLast page
1.
An efficient iterative approach to explainable feature learning
Dino Vlahek, Domen Mongus, 2023, original scientific article

Keywords: data classification, explainable artificial intelligence, feature learning, knowledge discovery
Published in DKUM: 13.06.2024; Views: 38; Downloads: 2
.pdf Full text (1,95 MB)
This document has many files! More...

2.
FairBoost: Boosting supervised learning for learning on multiple sensitive features
Ivona Colakovic, Sašo Karakatič, 2023, original scientific article

Abstract: The vast majority of machine learning research focuses on improving the correctness of the outcomes (i.e., accuracy, error-rate, and other metrics). However, the negative impact of machine learning outcomes can be substantial if the consequences marginalize certain groups of data, especially if certain groups of people are the ones being discriminated against. Thus, recent papers try to tackle the unfair treatment of certain groups of data (humans), but mostly focus on only one sensitive feature with binary values. In this paper, we propose an ensemble boosting FairBoost that takes into consideration fairness as well as accuracy to mitigate unfairness in classification tasks during the model training process. This method tries to close the gap between proposed approaches and real-world applications, where there is often more than one sensitive feature that contains multiple categories. The proposed approach checks the bias and corrects it through the iteration of building the boosted ensemble. The proposed FairBoost is tested within the experimental setting and compared to similar existing algorithms. The results on different datasets and settings show no significant changes in the overall quality of classification, while the fairness of the outcomes is vastly improved.
Keywords: fairness, boosting, machine learning, supervised learning
Published in DKUM: 11.06.2024; Views: 47; Downloads: 2
.pdf Full text (1,70 MB)
This document has many files! More...

3.
Gebrauch der Online-Tools und Apps beim DaF-Unterricht : Magisterarbeit
Ana Vok, 2024, master's thesis

Abstract: Die Magisterarbeit mit dem Titel „Gebrauch der Online-Tools und Apps beim DaF-Unterricht“ ist in einen theoretischen und einen empirischen Teil aufgeteilt. Im theoretischen Teil wird auf der Grundlage der analysierten Quellen die Digitalisierung und ihr Einfluss während der Covid-19-Pandemie beschrieben. Des Weiteren werden die Folgen der Digitalisierung nach der Covid-19-Pandemie in den Schulen erläutert. Dabei ist es entscheidend, dass Lehrer über digitale Kompetenzen für den Einsatz digitaler Medien verfügen. Im Weiteren werden die Begriffe M- und E-Learning mit den Online-Tools und Apps verbunden. Diese sind wichtig für das Fremdsprachenlernen, weshalb der Fokus auf dem DaF-Unterricht liegt. Es wird ermittelt, welche Kategorien von Online-Tools und Apps es gibt und wie sie in den DaF-Unterricht integriert werden können. Dabei wird erläutert, welche Kriterien für den Gebrauch von Online-Tools und Apps Lehrer beim Unterricht berücksichtigt werden müssen und welche Vor- und Nachteile beachtet werden sollen. Im empirischen Teil erfolgt die Analyse zweier Umfragen, an denen 81 Deutschlehrer und 127 Mittelschüler teilnahmen. Im Mittelpunkt steht die Analyse der Kompetenzen, der Häufigkeit und die Bereiche der Gebrauchs sowie die Vorteile und Hindernisse.
Keywords: DaF-Unterricht, digitale Medien, Online-Tools, Apps, M-Learning, E-Learning
Published in DKUM: 10.06.2024; Views: 35; Downloads: 3
.pdf Full text (2,07 MB)

4.
5.
A review of federated learning in agriculture
Krista Rizman Žalik, Mitja Žalik, 2023, review article

Keywords: federated learning, agriculture, architecture, data partitioning, federation scal, aggregation algorithms, communication bottleneck
Published in DKUM: 05.06.2024; Views: 63; Downloads: 5
.pdf Full text (839,33 KB)
This document has many files! More...

6.
Differences in self-regulated learning between gifted students, students with special needs and other students in Slovenian schools
Eva Kranjec, Karin Bakračevič, 2023, original scientific article

Abstract: Self-regulated learning strategies play a crucial role in learning progress and academic achievement of different groups of students. The purpose of the present study is to investigate differences in the use of self-regulatory strategies among a sample of 1,495 students, aged 12 to 15 years, representing three groups: gifted students, students with special needs, and other students. The theoretical framework for the study is Pintrich's (1991) model of self-regulated learning. Data were collected using the Motivated Strategies for Learning Questionnaire (MSLQ). Results indicated that gifted students scored significantly higher on the MSLQ subscales of motivation and learning strategies than students with special needs and other peers. Special needs students reported lower intrinsic and extrinsic goal orientation and weaker self-efficacy in learning and achievement than other students. There were no significant differences between these two groups on the MSLQ learning strategies subscales. Positive and statistically significant associations between the MSLQ subscales and final grades in three school subjects (Slovenian, mathematics, and foreign language) were also confirmed. We discuss the implications of our findings for future research and the educational context that contributes most to the development of self-regulated learning in all groups of students.
Keywords: self-regulated learning, MSLQ, gifted students, special needs, academic performance
Published in DKUM: 31.05.2024; Views: 83; Downloads: 0
.pdf Full text (514,48 KB)

7.
Machine learning driven extended matrix norm method for the solution of large-scale zero-sum matrix games
Burhaneddin İzgi, Murat Özkaya, Nazım Kemal Üre, Matjaž Perc, 2023, original scientific article

Abstract: In this paper, we develop a novel machine learning-driven framework for solving large-scale zero-sum matrix games by exploiting patterns discovered from the offline extended matrix norm method. Modern game theoretic tools such as the extended matrix norm method allow rapid estimation of the game values for small-scale zero-sum games by computing norms of the payoff matrix. However, as the number of strategies in the game increases, obtaining an accurate value estimation through the extended matrix norm method becomes more difficult. In this work, we propose a novel neural network architecture for large-scale zero-sum matrix games, which takes the estimations of the extended matrix norm method and payoff matrix as inputs, and provides a rapid estimation of the game value as the output. The proposed architecture is trained over various random zero-sum games of different dimensions. Results show that the developed framework can obtain accurate value predictions, with a less than 10% absolute relative error, for games with up to 50 strategies. Also of note, after the network is trained, solution predictions can be obtained in real-time, which makes the proposed method particularly useful for real-world applications.
Keywords: machine learning, EMN method, large-scale games, zero-sum games, approximated solutions
Published in DKUM: 31.05.2024; Views: 47; Downloads: 0
.pdf Full text (1,32 MB)
This document has many files! More...

8.
Changes in online distance learning behaviour of university students during the coronavirus disease 2019 outbreak, and development of the model of forced distance online learning preferences
Mateja Ploj Virtič, Kosta Dolenc, Andrej Šorgo, 2021, original scientific article

Abstract: Because of the Coronavirus Disease 2019 (COVID-19) outbreak, most universities were forced to choose Online Distance Learning (ODL). The study aimed to examine the response of university students to the new situation. A questionnaire was sent to the entire university student population. Based on responses from 606 students, it was revealed that use of all applications in ODL increased. However, only the use of MS Teams increased significantly, while the use of the other applications (email, Moodle, e-textbooks) increased in a range of low to medium in terms of effect sizes, and even nonsignificant for applications such as Padlet and Kahoot. Based on the replies of 414 respondents, a Model of Forced Distance Online Learning Preferences (MoFDOLP) based on Structural Equation Modeling was developed. With a chosen combination of predictors, we succeeded in predicting 95% of variance for Satisfaction, more than 50% for Continuance Preferences variance in MS Teams applications, and nearly 20% in the case of e-materials. Among hypothesized constructs, only Attitudes are a strong predictor of Satisfaction, while Organizational Support, Perceived Ease of Use and Learner Attitude toward Online Learning are not. Satisfaction is a good predictor of Continuance Preferences to use Information Technology after the lockdown ended.
Keywords: higher education, online distance learning, continuance preferences, COVID-19
Published in DKUM: 30.05.2024; Views: 35; Downloads: 0
.pdf Full text (1,26 MB)
This document has many files! More...

9.
Principal component analysis and manifold learning techniques for the design of brain-computer interfaces based on steady-state visually evoked potentials
Bartu Yesilkaya, Ebru Sayilgan, Yilmaz Kemal Yuce, Matjaž Perc, Yalcin Isler, 2023, original scientific article

Abstract: Steady-state visually evoked potentials (SSVEP) are stochastic and nonstationary bioelectric signals. Because of these properties, it is difficult to achieve high classification accuracy, especially when many considered features lead to a complex structure. We therefore propose a manifold learning framework to decrease the number of features and to classify SSVEP data by comparing lower dimensional matrices with well-known machine learning algorithms. We use the AVI-SSVEP Dataset, which includes stimuli at seven different frequencies and 15360 samples per person. The SSVEP features are extracted from relevant and distinctive frequency-domain, time-domain, and time–frequency domain properties, creating a total of 55 feature vectors. We then analyze and compare five divergent manifold learning methods with respect to their performance on nine different machine-learning algorithms. Among all considered manifold learning methods, we show that the Principal Component Analysis has the best classifier performance with an average of 22 components. Moreover, the Naive Bayes classifier with the Principal Component Analysis achieves the maximum accuracy of 50.0%–80.95% for a 7-class classification problem. Our research thus shows that the proposed analytical framework can significantly improve the decoding accuracy of 7-class SSVEP problems, and that it exhibits notable robustness and efficiency for small group datasets.
Keywords: manifold learning, brain-computer interface, steady-state visual evoked potential, principal component analysis, feature reduction
Published in DKUM: 29.05.2024; Views: 86; Downloads: 0
.pdf Full text (816,82 KB)
This document has many files! More...

10.
Machine learning partners in criminal networks
Diego D. Lopes, Bruno R. da Cunha, Alvaro F. Martins, Sebastián Gonçalves, Ervin K. Lenzi, Quentin S. Hanley, Matjaž Perc, Haroldo V. Ribeiro, 2022, original scientific article

Abstract: Recent research has shown that criminal networks have complex organizational structures, but whether this can be used to predict static and dynamic properties of criminal networks remains little explored. Here, by combining graph representation learning and machine learning methods, we show that structural properties of political corruption, police intelligence, and money laundering networks can be used to recover missing criminal partnerships, distinguish among diferent types of criminal and legal associations, as well as predict the total amount of money exchanged among criminal agents, all with outstanding accuracy. We also show that our approach can anticipate future criminal associations during the dynamic growth of corruption networks with signifcant accuracy. Thus, similar to evidence found at crime scenes, we conclude that structural patterns of criminal networks carry crucial information about illegal activities, which allows machine learning methods to predict missing information and even anticipate future criminal behavior.
Keywords: machine learning, crime, network, social physics
Published in DKUM: 28.05.2024; Views: 562; Downloads: 0

Search done in 0.28 sec.
Back to top
Logos of partners University of Maribor University of Ljubljana University of Primorska University of Nova Gorica