Search the repository

A+ | A- | | SLO | ENG

Abstract: Retailers traditionally rely on generic promotional catalogues that provide identical offers to all customers, regardless of their purchase histories or preferences. This “one-size-fits-all” approach often results in low engagement and inefficient allocation of marketing resources. This study investigates how cloud computing and artificial intelligence (AI) can automate the generation of personalized promotional offers, improving both customer relevance and retailer efficiency. Using a mixed-methods design science approach, requirements were gathered through interviews with marketing and IT staff, and a recommender system was developed by integrating ERP data, transaction histories, and product attributes into a private cloud environment. From a weekly pool of over 150 promotions, including more than 600 products in promotion, the system generated personalized lists of 6–20 promotions per customer, delivered via email, mobile applications, and print-ready flyers. A 12-week randomized controlled trial involving 400,000 loyalty program members evaluated the solution’s impact. Results show that customers receiving personalized offers generated 10% higher sales compared with the control group. Moreover, a positive correlation was observed between exposure frequency and both shopping frequency and basket growth, ranging from 6% to 36%. Marketing processes were standardized, improving campaign quality and consistency while reducing manual variability. Qualitative feedback confirmed that customers perceived the offers as more relevant and convenient, while highlighting the importance of clear communication about data usage, even though privacy is formally managed through the loyalty program. This study provides empirical evidence of the measurable business impact of hybrid recommender systems in grocery retail and emphasizes the importance of transparency, governance, and trust for successful AI-driven personalization.
Keywords: artificial intelligence, cloud computing, grocery retail, personalization, recommender systems
Published in ReVIS: 22.01.2026; Views: 557; Downloads: 2
Full text (12,52 MB)

Abstract: The use of artificial intelligence (AI) in education is a topic of growing research interest as its application in schools and universities is growing exponentially. This article addresses five research questions: current trends and statistics on the use of AI in education, the advantages and disadvantages of its use, the different perspectives of students and teachers, AI-based tools in education, and how AI could change education in the coming years. The findings highlight a significant growth in the use of AI, from adaptive learning platforms and automated assessment tools to personalised teaching systems. While AI improves efficiency, personalisation, and accessibility, it also raises concerns about data privacy, equity, and over-reliance on technology. A comparative theoretical analysis shows that students often emphasise the convenience and benefits of AI for inclusion, while teachers are more cautious and focus on pedagogical effectiveness and ethical issues. An overview of the main categories of AI tools used in education includes tools for people with disabilities, intelligent tutors and learning agents, chatbots, personalised learning systems, and visualisations and virtual reality. In the future, AI is expected to influence policy debates on curriculum design, teacher training, data management, and digital inclusion. By exploring the opportunities and challenges, while also connections, contradictions, and gaps in the literature, this study advises educators and students to use AI cautiously, informedly, and responsibly, ensuring that artificial intelligence supports rather than replaces human-centred teaching and learning.
Keywords: artificial intelligence in education, transformation of education, teaching and learning
Published in ReVIS: 22.01.2026; Views: 409; Downloads: 5
Full text (12,52 MB)

Abstract: Quality is often regarded as an abstract and elusive concept, becoming tangible only when deficiencies undermine real-world decisions. In spatial planning, for example, inaccuracies in planned land use data within the building permit process may lead to biased outcomes or, in critical cases, impede decision-making altogether. This study examines use cases of spatial data quality issues and their potential consequences, emphasizing both technical dimensions and the human factor, with particular attention to the gap between perceived and actual data quality. The research introduces the concept of this discrepancy as a key challenge in governance, with implications for decisionmaking processes and stakeholder trust. Building on these insights, an innovative approach to spatial data quality management is proposed through an automated Data Quality Management (DQM) framework. The framework is operationalized via the OPIAvalid toolkit, designed to enhance the effectiveness and reliability of national Spatial Information Systems (PIS). The expected contribution lies in providing a systematic methodology for managing spatial data quality, thereby strengthening evidence-based decision-making and fostering greater trust in spatial data infrastructures.
Keywords: perceived and actual data quality, data quality management (DQM), quality assurance/quality control (QA/QC), spatial data quality, automation
Published in ReVIS: 22.01.2026; Views: 384; Downloads: 2
Full text (12,52 MB)

Abstract: In the presentation, we address the issue of loneliness in relation to the increasingly all-encompassing use of the AI and the role reflexivity is playing in this regard. Based on previous research we presume that reflexivity significantly affects how one constitutes the world, engages in social relationship, navigates the media contents, and also engages in purposive action related to a more sustainable future. Based on that, we hypothesise that those who are highly reflexive tend to better cope with the potentially negative effects of AI – maintaining healthier social relations and personal empowerment. Research has been conducted on a national representative sample in Slovenia enabling a comparison between the older generations and the younger ones that have experienced direct and immediate socialization in a technologically mediated world, increasingly supported by forms of Artificial Intelligence
Keywords: loneliness, reflexivity, AI in youth
Published in ReVIS: 22.01.2026; Views: 371; Downloads: 3
Full text (12,52 MB)

Abstract: In this article we discuss how should we benchmark community detection algorithms in complex networks. We compare the community detection algorithms Louvain, Leiden, Label Propagation, Fast Label Propagation, Greedy modularity, Infomap, Walktrap and Girvan-Newman on complex networks of the Zachary karate club, Synthetic Network, a social network from X (Twitter), a neuroscience network, a email network and a patent citation network in the USA. We find that the speed of algorithms depends on the size and structure of networks. It turns out that among the considered algorithms for community detection in large networks, the Leiden algorithm is the most suitable, while on average the Fast Label Propagation algorithm performed the fastest in all cases. It is shown that on LFR benchmark network, algorithms successfully detect the same number of communities, however when we apply the same algorithms on complex networks the results are variable based on specific algorithm.
Keywords: community detection, networks and graphs, network analysis, complex networks
Published in ReVIS: 22.01.2026; Views: 409; Downloads: 7
Full text (12,52 MB)