doi.mendelu.cz: 28th International Conference Economic Competitiveness and Sustainability 2026

ISBN online: 978-80-7701-102-0 | DOI: 10.11118/ecos2026

28th International Conference Economic Competitiveness and Sustainability 2026

26.–27. 4. 2026 Brno

Petr David (ed.), Hana Vránová (ed.)

Ve dnech 26. a 27. března 2026 uspořádala Provozně ekonomická fakulta Mendelovy univerzity v Brně již 28. ročník mezinárodní konference „Economic Competitiveness and Sustainability“, kterého se zúčastnilo přes 100 účastníků z České republiky, Albánie, Francie, Německa, Maďarska, Litvy, Malajsie, Polska, Srbska, Slovenska, Ukrajiny a USA. V předloženém sborníku najdete 34 příspěvků, které byly doporučeny diskutanty konference a vybrány na základě recenzního řízení. Prezentované výstupy výzkumu přispívají a rozšiřují současný stav poznání a podněcují další diskusi nejen na akademické půdě, ale i v rámci veřejného a soukromého sektoru. Publikované příspěvky reflektují současné interdisciplinární výzvy vznikající na průsečíku digitální transformace, umělé inteligence, udržitelnosti, podnikových informačních systémů, ! financí, veřejné správy a hospodářské politiky, marketingu a managementu. Současně poukazují na rostoucí význam datově podloženého rozhodování, inteligentních technologií a udržitelných řešení pro zvládání aktuálních ekonomických a společenských výzev.

online: 2026, vydavatel: Mendel University in Brno

Příspěvky konference

ERP System Security Through the Eyes of Users: Evidence from Serbian Enterprises

Sanela Arsić, Aleksandra Radić, Đorđe Nikolić

DOI: 10.11118/ecos2026_001

Enterprise resource planning (ERP) systems are integrated business platforms characterised by a high need for security, which is a fundamental requirement for efficient operation. Although many vendors include comprehensive security mechanisms in their ERP solutions, user perceptions of security are also crucial for adequate system adoption. This study examines users’ perceptions of ERP system security and reliability in relation to demographic characteristics. Analysis of Variance (ANOVA) was used to identify differences in responses. The empirical research included a sample of 206 ERP system users from organisations in Serbia. The ANOVA results show that user perceptions differ significantly by users’ level of education, the organisation’s industry sector and economic activity, and the frequency and duration of system use. These findings provide valuable insights into variations in perceptions of ERP system security and help identify behavioural patterns influenced by demographic factors.

CIRCULAR ECONOMY INDICATORS AND DIGITAL TRANSFORMATION: ASSESSING THE RELATIONSHIP BETWEEN ERP ADOPTION AND CIRCULAR ECONOMY PERFORMANCE ACROSS EU COUNTRIES

Sanela Arsić, Aleksandra Radić, Anđelka Stojanović

DOI: 10.11118/ecos2026_002

The shift towards a circular economy requires the systematic integration of sustainability principles into economic and business processes, with digital technologies playing an increasingly important role. Among these, enterprise resource planning (ERP) systems represent an important tool, as they enable the collection, integration, and analysis of data crucial for enhancing resource efficiency and implementing circular economy principles. This study aims to examine the relationship between enterprise-level digitalization, ERP system adoption, and circular economy performance at the country level by using circular economy (CE) indicators across European Union member states. The research adopts a two-phase methodological approach. In the first phase, the TOPSIS multi-criteria decision-making method was applied to rank EU member states based on a set of relevant CE indicators. In the second phase, the resulting TOPSIS scores were used to examine the relationships among circular economy performance, digitalisation, and ERP system adoption using regression analysis and a group-based comparative technique. The results confirm pronounced differences in the performance of the circular economy across the member states of the European Union, with innovative and industrially advanced countries occupying leading positions in the TOPSIS ranking. At the same time, the findings indicate that digitisation at the enterprise level, including the adoption of ERP systems and ICT technologies, is positively associated with circular economy outcomes at the macro-country level, but by itself is not sufficient to achieve high performance without adequate structural and innovation capacities. The findings should be interpreted as macro-level associations rather than direct causal effects, as the study relies on aggregated country-level indicators and proxy measures of enterprise digitalisation.

ECONOMETRIC MODEL OF GENERALIZED LINEAR REGRESSION FOR PREDICTING FUTURE NUMBERS OF INSURED PEOPLE BY GENERAL HEALTH INSURANCE COMPANY OF THE CZECH REPUBLIC – COMPARISON OF RESULTS AFTER FIVE YEARS

Jaromír Běláček

DOI: 10.11118/ecos2026_003

This paper compares the results of predicting the number of insured persons of the General Health Insurance Company of the Czech Republic in 2018, based on a generalized-discounted linear regression model (GLM), which formalizes the projected number of insured persons in age groups through exogenous demographic, re-reregistration, and formal linear and quadratic trend components, with reality observed in 2018–2022. Compared to standard in-sample methods, derived population projections, or the traditional indexing method, the 2018 method, where the calibration of discount factors α meets the criteria for maximizing the determination coefficient R2(α), overall yields better results. Accurate predictions of future numbers of insured individuals have the potential to better calculate and plan future expenditures of the health insurance funds during current period of population aging, and for several years ahead, too.

OPEN-SOURCE DECEPTION SECURITY FOR AWS CLOUD ENVIRONMENTS: A COST-CONSCIOUS ARCHITECTURE FOR SMES

Matěj Cajthaml, Tomáš Pitner

DOI: 10.11118/ecos2026_004

This paper presents the design and verification of an open-source deception security solution for Amazon Web Services cloud environments with a focus on small and medium-sized businesses and other budget-constrained organizational units. The study starts from the practical problem that commercial deception platforms provide centralized management, automated deployment, alerting, and visualization, yet their licensing and operational costs are often difficult to justify for smaller organizations. At the same time, cloud adoption continues to grow and expands the attack surface exposed through identities, storage services, virtual machines, and application interfaces. The proposed approach combines infrastructure as code, deceptive assets, and a lightweight monitoring pipeline to provide early warning signals without requiring a costly proprietary stack. The architecture integrates decoy cloud resources, identity-based honeytokens, event filtering, message queuing, log forwarding, and visualization components. The implementation was validated through a simulated insider-oriented attack path mapped to selected MITRE ATT&CK techniques. The results indicate that a reproducible and operationally realistic deception layer can be built from open-source tools and native cloud services while preserving constrained permissions and manageable running costs. The paper concludes that open-source deception can serve as a practical complementary layer in defense-in-depth strategies for cloud environments used by smaller organizations.

AN OPEN ETL AND RAG PIPELINE FOR UNIFIED ACCESS TO HETEROGENEOUS CLIMATE DATA

Dávid Chomo, Jaromír Landa

DOI: 10.11118/ecos2026_005

Climate data are dispersed across numerous portals in incompatible formats, presenting a substantial barrier for researchers and analysts working on climate change and sustainability. This paper presents an open system that automates the processing of climate datasets end-to-end. The system covers the entire workflow, from the downloading of raw files through to format detection (NetCDF, GRIB, GeoTIFF, CSV, HDF5, and ZIP archives), and transformation into a text description with statistical characteristics. The system also creates vector embeddings and natural language querying. The ETL pipeline is orchestrated by Dagster, the vector store is provided by Qdrant with an HNSW index, and response generation takes place in a multi-level retrieval-augmented generation (RAG) pipeline combining weighted Reciprocal Rank Fusion, Maximal Marginal Relevance diversification, and a Corrective RAG validation layer. The integration of a web interface, constructed on the foundations of Vue.js 3 and FastAPI, serves to facilitate the accessibility of this functionality to non-technical users. The system achieved an average composite score of 91% on a set of ten queries, with all tests passing the 80% threshold. Following a series of optimizations, the average query latency decreased from 120 to 180 seconds to approximately 26 seconds. The primary contribution of this work lies in the integration of components that currently function independently within existing climate RAG systems, thereby reducing the barriers to accessing standardized climate data for sustainability research.

SOCIAL ENGINEERING AND HUMAN-FACTOR VULNERABILITIES IN CRITICAL INFRASTRUCTURE: FINDINGS FROM A SIMULATED PHISHING CAMPAIGN

Kateřina Dobrovolná, Tomáš Pitner

DOI: 10.11118/ecos2026_006

Critical infrastructure organizations increasingly face cyberattacks that exploit human behaviour rather than technical vulnerabilities. Social engineering attacks, especially phishing and spear-phishing, remain among the most effective attack vectors because they target employees directly. This paper analyses the role of the human factor in cybersecurity within a critical infrastructure environment through a combination of threat intelligence analysis, questionnaire-based research, and a simulated phishing campaign implemented using the GoPhish framework. The study evaluates employee awareness, behavioural vulnerabilities, and organizational preparedness against social engineering attacks. The results confirm that phishing remains one of the dominant cyber threats and that technical protection alone is insufficient without systematic employee education and awareness-building. The paper proposes a set of organizational and technical countermeasures focused on strengthening the organizational “human firewall.” The findings contribute to current discussions on cybersecurity resilience in critical infrastructure and highlight the importance of integrating continuous awareness programs with practical simulation exercises.

TRUST BETWEEN INDIVIDUALS: AN EMPIRICAL ANALYSIS INSPIRED BY COVEY AND SINEK

Tereza Frömmelová, Kateřina Pojkarová

DOI: 10.11118/ecos2026_007

This paper examines trust as described in the works of Simon Sinek and Stephen M. R. Covey, which together form the theoretical framework for this study. Their approaches view trust as a relational and value-based process that develops through consistent behavior, integrity, and psychological safety. Based on this theoretical framework, the authors of the paper conducted a questionnaire survey among university students in 2025. The study aimed to explore how young adults understand trust, how they apply it in personal and team relationships, and whether their views support or challenge the concepts proposed by Covey and Sinek. The findings show that students’ views on trust in teamwork reflect many elements of contemporary trust theory, particularly the emphasis on authenticity, reliability, and psychological safety. The paper contributes to a deeper understanding of how theoretical models of trust are translated into lived experiences among young adults.

FROM VALUES TO ACTION: EXPLAINING THE INTENTION-BEHAVIOR GAP IN GREEN FASHION CONSUMPTION AMONG GENERATION Z IN BANGLADESH

Shovan Nasim Haider, Miloslava Chovancova

DOI: 10.11118/ecos2026_008

This study aims to examine how green consumer values are translated into actual green fashion apparel purchasing behavior among Generation Z consumers in Bangladesh, with particular focus on explaining the persistent intention–behavior gap within a developing, apparel-producing economy. The study adopts a quantitative research design grounded in a value–intention–behavior framework. Drawing on the Theory of Consumption Values and the Theory of Planned Behavior, the proposed conceptual model incorporates green purchase intention as a central mediating mechanism, supported by post-intention constructs including implementation intention and differentiated self-efficacy (action and coping). Survey data are collected from Generation Z consumers in Bangladesh and analyzed using partial least squares structural equation modeling (PLS-SEM) to examine direct, mediated, and moderated relationships among the constructs. Building on the theoretical and empirical gaps identified in the literature, the study focuses on clarifying the mechanisms through which multidimensional consumer values influence green purchase intention and how intentions are subsequently enacted under contextual constraints. Particular attention is given to the roles of implementation intention and self-efficacy in strengthening intention–behavior consistency in sustainable fashion consumption. By moving beyond intention-centric explanations, this study is positioned to advance theoretical understanding of green consumption behavior and provide contextually grounded insights relevant to developing producer economies. The findings are expected to inform future research and support practitioners and policymakers in designing interventions that facilitate the translation of sustainability values into actual purchasing behavior among young consumers.

ROBUSTNESS OF A FUZZY CONTROLLER VERSUS CLASSICAL PID CONTROL ON TYPICAL PLANTS

Marek Hebelka, Radim Farana

DOI: 10.11118/ecos2026_009

This article deals with the robustness comparison of a Mamdani-type fuzzy controller and a classical PID controller on four typical controlled systems: proportional non-oscillatory, proportional oscillatory, proportional with transport delay, and integrating with inertia. The work responds to the lack of systematic comparison of both controller types under disturbances and parameter changes. Measurements were performed in MATLAB Simulink; PID was tuned automatically by the PID Tuner, the fuzzy controller manually by the Mamdani method with a simplified 3×3 rule base. Robustness was assessed for a position step of 10–40% and time-constant changes of 0–20%. Control quality was quantified by the IAE and ITAE summation criteria. For three of the four systems, the fuzzy controller was one to two orders of magnitude better in ITAE; for the system with transport delay, the PID was conversely two orders of magnitude better.

PREDICTIVE ANALYTICS IN NEPALESE AGRI-CULTURE

Sandesh Kumar Kafle, František Dařena

DOI: 10.11118/ecos2026_010

This study addresses climate vulnerability and the lack of data-driven tools in Nepalese agriculture by building predictive crop-yield models for 14 districts in the Koshi Province. Applying a data-science pipeline to historical agricultural records and NASA POWER agroclimatic data spanning 11 crops over 11 fiscal years (2011/12–2021/22), the research evaluates Random Forest and Gradient Boosting Regressors across five distinct weather-feature datasets. The results demonstrate that the Random Forest Regressor consistently outperforms Gradient Boosting, explaining 74% to 84% of yield variance. The highest predictive accuracy was achieved using engineered growing-season averages (Baseline R2 = 0.83, Post-tuning R2 = 0.84), confirming that crop-specific seasonal summaries are the most informative features. Temperature emerged as the strongest single predictor of crop yield (R^2 = 0.82), while standalone precipitation exhibited the weakest predictive power (R2 = 0.74). These findings offer an empirical decision-support foundation for farmers and policymakers to quantify climate impacts and enhance sustainable agricultural planning in Nepal.

VISUAL ATTENTION TO INFLUENCER MARKETING CONTENT IN DECORATIVE COSMETICS: EVIDENCE FROM AN EYE-TRACKING EXPERIMENT

Kateřina Klátilová, Stanislav Mokrý, Jana Turčínková, Michal Pšurný

DOI: 10.11118/ecos2026_011

This study examines how selected elements of influencer marketing content affect consumer attention and perception in the context of decorative cosmetics. Primary data were collected through an eye-tracking experiment and follow-up in-depth interviews with 35 women aged 18 to 34. The experiment used visual and audiovisual stimuli based on social media content and included several A/B comparisons, such as user comments, number of hashtags, product arrangement, product photography style, make-up application, and the use of a well-known person’s face. The results suggest that product arrangement and photographic style influenced visual attention, while celebrity presence and make-up application did not lead to statistically significant differences in observation time. Interviews showed that participants generally perceived influencer reviews positively, especially when they appeared authentic, critical, and based on real product use. Most participants also preferred influencers to disclose paid cooperation.

A TOOL FOR AUTOMATICALLY TRANSCRIBING IN-TERVIEWS WITH SPEAKER RECOGNITION

Lukáš Kleveta, Jaromír Landa

DOI: 10.11118/ecos2026_012

The present paper presents the design and implementation of a software tool for automatically transcribing audio recordings, supplemented by speaker recognition (diarization). The tool integrates two open-source components: The implementation of OpenAI’s Whisper AI model for speech-to-text conversion and the pyannote.audio library-or, alternatively, NVIDIA NeMo - for speaker recognition. The architecture of the system is implemented as a linear pipeline with modules for audio preprocessing, diarization, transcription, and result merging. It is possible to modify key processing parameters (model, language, thresholds, clustering method) via a configuration file without modifying the code. The functionality was verified on two test sets. The first was a short English recording with a single speaker. In this case, a quantitative comparison of Whisper AI models was made using the Levenshtein metric. The second was a one-minute recording with two speakers. In this case, a comparison was made of twenty configurations in terms of transcription, diarization, and computational complexity. The medium and large-v2 models demonstrated the highest degree of mutual agreement among transcriptions (97.79%), while the lowest was exhibited by the tiny and large-v3 models (82.88%). With regard to overall balance, the configuration employing the large-v2 model and the NeMo diarization module with spectral clustering was identified as the optimal solution. The findings substantiate the practical viability of open-source instruments for automated transcription. However, they concurrently underscore the constraints associated with multilingual recordings, the computational intricacy, and the susceptibility of diarization to variations in audio quality.

INTEGRATION OF CHRONOS HIGH-SPEED CAMERAS INTO THE MERCURYRT DIGITAL IMAGE CORRELATION SYSTEM

Martin Kmenta

DOI: 10.11118/ecos2026_013

This article deals with the design and implementation of an integration module that connects Chronos 1.4 and 2.1-HD high-speed cameras with the MercuryRT digital im-age correlation (DIC) system. The work responds to the previously missing integration that required manual transfer of recordings and separate setup of both systems. The proposed solution uses the camera REST API for control and an RTSP stream decoded by an external FFmpeg process for live preview. A custom integer-based YUV420p-to-RGB conversion algorithm in C# was implemented; it handles a 10 FPS preview in real time with typical latency around 300 ms. Functionality and stability were verified by a four-hour stress test and trigger measurements using an LED flash; the recording completed without frame loss.

WEB APPLICATION FOR VISUALIZATION AND COMPARISON OF MINIMUM-REDUNDANCY CODES

Ivan Kolesnychenko, Radim Farana

DOI: 10.11118/ecos2026_014

This article presents the results of the development and testing of a specialized web application designed for teaching and comparing lossless data compression algorithms based on minimum redundancy codes. The work focuses on two fundamental pillars of statistical coding: the Shannon-Fano (Shannon 1948) and Huffman algorithms (Huffman 1952). The main contribution of the created solution is an interactive environment that allows users to monitor calculation steps and the creation of tree structures in real-time on identical inputs, which supports an intuitive understanding of the differences in the efficiency of both methods. The technical solution is built on modern Vue 3 and D3.js technologies. Validation took place in the form of user testing, which proved the high didactic potential of the application and its benefit for students of technical fields.

ONLINE PLATFORMS AS TOOLS FOR ENVIRONMENTAL CONSUMER EDUCATION

Lívia Kolibárová, Irena Baláková

DOI: 10.11118/ecos2026_015

In the digital era, online platforms are becoming a significant channel for environmental consumer education. This paper aims to identify consumers’ preferred options for environmental education through online platforms, evaluate their actual utilization, and describe the motivations and barriers to engagement. The empirical section is based on the quantitative questionnaire survey primary data. To statistically verify relations, Spearman’s rank correlation coefficient and Pearson’s χ²-test were applied; furthermore, factor and cluster analyses were conducted. The results indicate that 92% of respondents utilize online platforms to obtain environmental information, with social media, and online articles being predominant. A weak positive correlation (ρ = 0,29) was found between the perceived availability of eco-friendly options in the place of residence and declared pro-environmental behavior, whereas a moderate positive correlation (ρ = 0,53) was demonstrated between the subjective level of environmental education and pro-environmental behavior. The preferences regarding platform types differ statistically significantly based on the respondents’ economic activity. Factor analysis identified three dimensions of attitudes toward online environmental education. The cluster analysis identified three distinct groups of respondents—passive online users, highly active and university-educated learners utilizing a wide spectrum of digital tools and disengaged non-users. The formulated recommendations are primarily directed at educational institutions and non-profit organizations.

CZECHUP: A WEB SYSTEM FOR PREPARING FOREIGNERS FOR THE CZECH LANGUAGE EXAM WITH ADAPTIVE VOCABULARY AND AI-GENERATED EXERCISES

Polina Komárková

DOI: 10.11118/ecos2026_016

This article deals with the design and implementation of the CzechUP web system that helps foreigners prepare for the mandatory Czech language exam required for permanent residence (level A2) or citizenship (level B1). The work responds to the lack of suitable tools for Russian- and Ukrainian-speaking users, for whom the most popular application Duolingo is available only via English. The system integrates a database of 4,372 words parsed from three official CEFR textbooks, the Majka morphological analyzer (up to 347 forms for a single adjective), exercise generation via the Gemini API, and translation via Seznam Slovník and DeepL. The backend is built in ASP.NET Core over PostgreSQL, the frontend in Next.js with localization into three languages. User testing with five respondents confirmed the usefulness of the grammatical-form filter as a unique feature and estimated the error rate of AI-generated exercises at approximately 10%.

GENERATION Z AND AI-DRIVEN SHOPPING: COMPARING ONLINE AND IN-STORE CONSUMER EXPERIENCES

Benjamin Křištof, Jana Dannhoferová

DOI: 10.11118/ecos2026_017

This paper examines the methods and intensity of Artificial Intelligence (AI) tool utilization by Generation Z within online and offline shopping environments. The empirical investigation employs a mixed-methods research design, combining a quantitative questionnaire survey (n = 114) and six semi-structured interviews conducted in April and May 2026 among respondents aged 18–27 in the Czech Republic. The objective was to identify dominant AI tools, key motivators, and barriers to their adoption, and to evaluate consumers‘ willingness to share personal data in both environments. Online interaction with AI is more conscious, focusing primarily on generative assistants (ChatGPT), visual search, and recommendation systems. In stores, self-checkout counters and interactive kiosks dominate, though their utilization is often unreflected by consumers. Statistical testing of five hypotheses revealed a higher willingness to share personal data in the online environment (p = 0.018). Surprisingly, however, it also demonstrated statistically significantly lower privacy concerns in the offline environment compared to online (p < 0.001), which contradicts the assumption derived from the literature. No gender differences in the frequency of AI usage were detected. Based on these findings, a “phygital” strategy was formulated, linking a unified customer data identity, the smartphone as a mediator between environments, and the principles of Explainable AI (XAI).

SERVICE MANAGEMENT FOR SMART BUILDING ADMINISTRATION USING VIRTUAL IOT SIMULATION AND PREDICTIVE ANALYTICS

David Kubart, Zuzana Špendel

DOI: 10.11118/ecos2026_018

This article addresses the design and implementation of a virtual service management model for smart building management within an enterprise environment. The work innovatively links the ITIL 4 methodology with a containerized IoT infrastructure, edge-cloud architecture, and predictive analytics. The objective of the research is to demonstrate that even a fully software-defined smart building environment can be systematically managed as a complex service with a clearly defined Service Level Agreement (SLA) and processes for effective incident and event management. To validate the proposed approach, a fully portable laboratory environment was developed using Docker containers, MQTT asynchronous communication, an InfluxDB telemetry data store, Grafana visualization, and Python modules for data generation and advanced analytics. Building behavior is dynamically configured via configuration files, allowing a single codebase to simulate various types of facilities with different operational priorities and SLA requirements. The proposed model was thoroughly verified through experimental scenarios, including simulations of network communication outages, MQTT message delays, sensor data errors, and complete physical disasters. The results demonstrate the practical usability and benefits of ITIL 4 principles (ITIL Foundation 2019) for digitalized building management, increasing data reliability, and automating operations.

LIGHTNING NETWORK AS AN ALTERNATIVE PAYMENT INFRASTRUCTURE FOR SMALL AND MEDIUM-SIZED ENTERPRISES

Tomáš Lacina, Petr Jedlička

DOI: 10.11118/ecos2026_019

This paper presents the design, implementation, and economic evaluation of a payment module that enables merchants to accept Bitcoin payments through the Lightning Network using a self-hosted Bitcoin node. While the Bitcoin base layer suffers from limited throughput, relatively high transaction fees, and longer confirmation times, the Lightning Network addresses these limitations through off-chain payment channels, enabling near-instant settlement at minimal cost. The study combines a theoretical analysis of the Bitcoin protocol and the Lightning Network with a comparative evaluation of traditional payment methods and cryptocurrency-based payments. Primary data were collected from the official pricing schedules of eight payment terminal providers and six payment gateway providers operating in the Czech Republic. A full Bitcoin and Lightning Network node was deployed on an HP 260 G2 Mini platform running Umbrel, Bitcoin Core, LND, and BTCPay Server. Based on this infrastructure, a web application named Flash was developed to support payment processing and transaction management. The functionality and usability of the application were verified through pilot user testing. The results indicate that transaction costs associated with the Lightning Network are several orders of magnitude lower than those of conventional payment terminals and payment gateways while maintaining immediate settlement. Furthermore, operating a self-hosted payment infrastructure requires only a modest initial investment. The study identifies merchant technical expertise and the limited adoption of Lightning Network wallets among consumers as the principal barriers to broader deployment. The findings suggest that the Lightning Network represents a technically feasible and economically attractive payment alternative for small and medium-sized enterprises seeking to reduce transaction costs and increase payment-processing autonomy.

FINE-TUNING ROBERTA FOR SENTIMENT ANALYSIS OF SUPPORT CONVERSATIONS WITH PRIVACY-PRESERVING ANONYMISATION

Vojtech Mahdal, Oldrich Trenz

DOI: 10.11118/ecos2026_020

This article describes a prototype system for processing support-conversation text. The system accepts a text input, classifies sentiment into three classes (positive, neutral, negative), masks selected sensitive items such as names, e-mails, phone numbers, URLs, usernames and long numeric strings, and stores an anonymised record. The sentiment classifier is based on RoBERTa and was fine-tuned in three stages on public chat and dialogue datasets. Evaluation showed that the second stage reached the best score on its internal test split, but its accuracy on a separate combined test set fell from 83.41% to 70.06%. After a final fine-tuning stage on a smaller dataset with a higher share of neutral examples, the model reached 92.82% accuracy on 1,573 independent samples. The anonymisation layer is rule-based and is most suitable for structured patterns; names, especially sentence-initial capitalised words, remain the main limitation. The result is a working demonstration consisting of a FastAPI service, a JavaScript dashboard and an SQLite log, containerised with Docker and deployed on Hugging Face Spaces.

INCOME, ATTITUDES, AND FOOD WASTE: SEGMENTATION OF CZECH HOUSEHOLDS

Anna Martinková, Veronika Antošová, Lucie Veselá

DOI: 10.11118/ecos2026_021

Household food waste is a significant environmental and economic problem in the Czech Republic — it is estimated that households generate a significant portion of total food waste. The present article examines whether household income affects the level of food waste and attitudes towards it, and complements this survey on consumer segmentation based on purchasing and consumption behavior. The empirical basis is a questionnaire survey conducted in January and February 2025 among the population of the Czech Republic (n = 294), supplemented by semi-structured in-depth interviews (n = 6). The statistical dependence between the monitored variables was verified by Pearson‘s chi-square test of independence, and the strength of the dependence was verified by Cramer‘s coefficient; hierarchical cluster analysis using the Ward method was used to segment respondents. The results showed that the proportion of food thrown away per month was not statistically dependent on income, while attitudes towards waste did depend on income and age. A significant predictor of lower waste was the frequency of checking the refrigerator. Cluster analysis identified three household segments with different purchasing, waste and perception approaches. The findings suggest that quantitatively measured waste levels and attitudes towards the problem may diverge, which has implications for the development of targeted communication strategies.

CONTROL OF INTEGRATING SYSTEMS USING A TWO-DEGREE-OF-FREEDOM PID CONTROLLER

Jakub Meduna, Radim Farana

DOI: 10.11118/ecos2026_022

This article deals with the comparison of a two-degree-of-freedom (2DOF) PID controller and the classical 1DOF PID controller for the control of integrating systems. The work responds to the structural limitation of the 1DOF controller, which cannot simultaneously optimize setpoint tracking and disturbance rejection. Three typical integrating plants (with transport delay, oscillatory, non-oscillatory) and four controller variants tuned by the SIMC method and the PID Tuner application were tested in the MATLAB Simulink environment. Control quality was assessed using the IAE and ITAE criteria. The serial 2DOF controller improved ITAE by 28 to 54% compared with the best 1DOF solution; the optimal tuning method, however, depends on the plant structure.

PRACTICAL BLOCKCHAIN DEPLOYMENT FOR TICKET MANAGEMENT: A CASE STUDY WITH SOLIDITY SMART CONTRACTS

Jan Mrákava, Radim Farana

DOI: 10.11118/ecos2026_023

This article deals with the design and implementation of a web application for ticket management based on a local Ethereum-compatible blockchain. The work responds to the gap between marketing claims about blockchain as a universal document verification tool and limited knowledge of real transaction costs. The frontend uses React.js, the backend Node.js with Express and MongoDB, and the logic consists of two Solidity 0.8.10 smart contracts (EventContract, TicketContract) deployed through Truffle into the Ganache environment (Chain ID 1337). Measurements showed that creating a single event with a typical description costs 0.0078 ETH (~546 CZK at the 2024 exchange rate), while ticket purchase and resale require an order of magnitude lower fees. The main finding: transaction cost is dominated by the length of text strings, which for document verification in general means storing only hashes and short identifiers on the chain.

GREENWASHING IN THE MARKETING COMMUNICATION OF SUSTAINABLE FASHION COMPANIES: THE ATTITUDES OF GENERATION Y AND Z

Gabriela Navrátilová, Veronika Antošová, Lucie Veselá

DOI: 10.11118/ecos2026_024

The paper deals with the perception of the credibility of environmental claims of fash-ion brands among Czech consumers of generation Y and Z and the possibilities of communication that minimizes the risk of greenwashing. The aim is to identify the key factors influencing the purchasing behaviour of these generations in the sustainable fashion market and to describe their attitudes towards the credibility of sustainability marketing messages. A combination of situational analysis of secondary data and a primary questionnaire survey in the period March–April 2024 is used. The sample con-sists of 368 respondents born between 1980 and 2010 selected according to their age. Dependencies are tested by Pearson’s chi-square independence test at a significance level of 5%, the intensity of association is measured by Cramér’s V. It was found that 68% of respondents are more likely to distrust marketing claims about the sustainabil-ity of fashion brands, and 65.5% say they would not buy from a brand convicted of greenwashing. Confidence in environmental claims is not statistically dependent on the age of the respondent or the frequency of purchases of slow fashion, but it is weakly but statistically significantly dependent on the knowledge of the concept of greenwash-ing and on the frequency of its feeling when buying clothes. Transparent communica-tion about the origin of materials and production processes increases consumer trust for 58.7% of respondents. The results support the conclusion that greenwashing risk reduction is based primarily on the transparency, authenticity and traceability of the message, not on the intensity of the communication itself.

INTERACTIVE TRANSFORMATION OF LAYERS IN 3D TILES FORMAT WITHIN A WEB GIS APPLICATION

Markéta Nedělová, Ivo Pisařovic

DOI: 10.11118/ecos2026_025

The paper deals with the design and implementation of an extension to the Spatial Lens web application, which is used to prepare 3D educational materials displayed in virtual reality. The subject of the extension is editing spatial layers stored in the 3D Tiles format, specifically their displacement, rotation and scaling directly in the scene preview, which eliminates the need to use external tools. Three-dimensional geodata is not always perfectly prepared for a specific scene and may have a different axis orientation or incorrect scale. The entire 4×4 transformation matrix is not transferred between the frontend and backend, but a more economical data structure containing the position (longitude, latitude and altitude), rotation and scale. This solution reduces cumulative rounding errors, increases accuracy for large numbers, and facilitates input validation. The backend recalculates the resulting matrix each time from the beginning from the saved original matrix and a stable pivot point, thus ensuring idempotence of repeated saves. The usability of the extension was verified through scenario-based user testing with ten participants. The results confirm the usability of the tool while identifying areas for further improvement of the user experience.

A WEB-BASED FINANCIAL PLANNING APPLICATION FOR ENTERPRISE BUDGETING AND INVESTMENT MANAGEMENT: A DESIGN AND IMPLEMENTATION CASE STUDY

Ondřej Pavlík, Oldřich Trenz

DOI: 10.11118/ecos2026_026

A mid-sized company that tripled its number of organizational units over five years outgrew its spreadsheet-based budgeting. Centre managers planned in their own Excel files, the finance department merged them by hand, and errors surfaced only after the fact. This paper reports the design and implementation of an on-premise web application that sits between the company ERP system and the centre managers, replacing the manual consolidation with a single validated planning workflow. The work follows a standard software-engineering process: analysis of the existing routine, requirements (22 functional, 12 non-functional), a make-or-buy assessment of three commercial vendors, UI design in Figma, implementation, testing, and deployment. The application is built on ASP.NET Core (.NET 8) with Razor Pages, a Microsoft SQL Server database, SOAP integration with the ERP, integrated Windows authentication, and Power BI reporting embedded through OAuth 2.0. None of the three vendors covered the on-premise requirement, customization, and integration at once, so a tailored solution was developed and is now in production. We describe the architecture, the integration and authentication decisions, and five concrete problems solved during development. No quantitative performance metrics were measured; the contribution is the engineering design.

AUTHORSHIP VERIFICATION OF ENGLISH TEXT: TRADITIONAL STYLOMETRY, TRANSFORMERS, AND LARGE LANGUAGE MODELS COMPARED AND DEPLOYED

Jiří Pokorný, Michaela Veselá, Oldřich Trenz

DOI: 10.11118/ecos2026_027

This article compares three families of methods for authorship verification, the binary decision whether two texts share one author. We frame the task as pairwise classification and test traditional stylometric pipelines (bag-of-words and TF-IDF with scikit-learn classifiers), fine-tuned transformers (Longformer, BigBird-RoBERTa, ModernBERT), and large language models used through prompting. Training used the PAN 2020 corpus, testing the disjoint PAN 2021 corpus, together 295,800 English text pairs. A fine-tuned Longformer classification head reached the highest F1 of 92.38%, but a much cheaper TF-IDF pipeline with a linear classifier trailed it by less than one percentage point at 91.50%. The way two text vectors are combined into one feature vector mattered more than the classifier itself: absolute difference beat element-wise multiplication everywhere. The three open large language models we tried produced unusable output. Because hosting was constrained, the deployed web application runs TF-IDF with logistic regression (F1 91.28%), which returns the probability the interface needs.

PRODUCT NAME MATCHING

Jan Přichystal

DOI: 10.11118/ecos2026_028

The problem of product name matching belongs to the area of entity matching which is one of the data integration tasks. It plays an important role in the analysis of data available on the internet and it combines various techniques and data preprocessing. Gathering and analyzing customer reviews about products published on the internet is challenging task, because particular e-shops offer the products under a bit different names and it could be complicated to match identical products. The article aims at selecting the identical products identified by their names and introduces methods for product name matching based on machine learning using the combination of token-based and character-based functions.

SUMMARY OF FINANCIAL ARTICLES

Jan Přichystal, Tomáš Jakúbek, Roman Valovič

DOI: 10.11118/ecos2026_029

The Summary of Financial Articles (SFA) application is intended for obtaining articles in the field of financial markets published by various online media and their subsequent analysis. The aim is to facilitate the work of financial analysts, who have to process a large amount of information on a daily basis, use it to create a picture of the state of financial markets and individual securities, and then issue recommendations to customers on their trading strategy. Gathering the articles and news from online sources is very demanding process due to the technologies used there and ever-changing structure of these web sites. Not only the web application but also the API for integration into customer applications is available.

NPL STUDIO

Jan Přichystal, Roman Valovič, Tomáš Jakúbek

DOI: 10.11118/ecos2026_030

This article deals with the description of NLP Studio software. This web application allows users to prepare their own textmining experiments for the natural language analysis without the need to use the programming language. There is a development environment in which the user can work in the form of visual programming. Preprocessing, training or visualization components are available. The software is designed primarily for teaching natural language processing and aims to help students navigate this area even if they do not have a deep knowledge of programming.

FACTORS INFLUENCING THE SELECTION OF CRAFT BEER: A FACTOR ANALYSIS AND GENERATIONAL SEGMENTATION IN THE CZECH MARKET

Lenka Procházková, Marek Záboj, Matěj Sysel

DOI: 10.11118/ecos2026_031

This article presents the results of an empirical study of consumer behavior in the Czech craft beer market. The primary objective was to identify the key factors influencing the selection of craft beer and, based on these factors, to perform a market segmentation analysis. The research was conducted through a quantitative survey combining online distribution via social media platforms and offline distribution using QR codes placed in craft beer pubs in Brno. After the application of screening criteria, a total of 458 respondents aged 18 to 57 years were included in the statistical analysis. Ten formulated hypotheses were tested using non-parametric statisti-cal methods (Pearson’s chi-square test, the Mann–Whitney U test, the Wilcoxon signed-rank test, and the Kruskal–Wallis test). The principal analytical method was exploratory factor analysis with Varimax rotation, which extracted four factors explaining 43.45% of the total variance: price sensitivity and conservatism, taste and sensory attributes of beer, authenticity and exclusivity, and social and visual influences. Taste and aroma emerged as the most influ-ential individual determinants of craft beer selection, whereas recommendations from influ-encers were found to be the least influential. Generational segmentation revealed that Genera-tion Y constitutes the most active market segment, while Generation Z consumes craft beer significantly less frequently. The findings suggest that value-based attributes, such as authen-ticity and regional origin, outweigh price sensitivity and traditional forms of advertising among craft beer consumers in the Czech Republic.

CLOSING THE GREEN GAP: HARMONIZING PROJECT AND PROCESS APPROACHES IN HIGHER EDUCATION VIA SUSTAINABLE LEADERSHIP AND DIGITAL INTERNAL MARKETING

Petra Skořepová, Petra Marešová

DOI: 10.11118/ecos2026_032

Higher Education Institutions (HEIs) are currently navigating the complexities of the global “Twin Transition” (green and digital). However, a significant “implementation gap” persists between strategic sustainability declarations and operational reality. This paper scrutinizes this phenomenon within the University of Hradec Králové (UHK), analyzing the misalignment between the university‘s “Sustainable Development Strategy 2030” and actual workforce perception. Drawing on the duality of project and process approaches in institutional adaptability, this study utilizes qualitative data from 15 in-depth semi-structured interviews. The findings delineate a critical asymmetry: while respondents widely recognize tangible infrastructure improvements (“hardware”) such as waste sorting stations and energy-saving technologies, they report a profound lack of awareness regarding the strategic vision. The analysis identifies key barriers preventing behavioral adoption, specifically “bureaucratic inertia” where digital goals clash with persistent analog administrative processes, and a “skill gap” where staff lack the know-how to operate new technologies (e.g., smart thermostats), leading to frustration and disengagement. The study argues that without addressing the behavioral dimension through specific internal marketing strategies, the potential of green infrastructure remains untapped, resulting in institutional decoupling. The paper concludes that leadership must pivot towards an Industry 5.0 mindset, leveraging digital tools not just for management, but as internal marketing instruments to “nudge” the sustainable workforce towards active participation.

CLASSIFYING SCANNING ELECTRON MICROSCOPE DEFECTS WITH A CNN AND A CUSTOM FOURIER-TRANSFORM LAYER

Magdaléna Skřičilová, Oldřich Trenz

DOI: 10.11118/ecos2026_033

A scanning electron microscope has to be checked periodically, and two of its most common image faults, defocus and astigmatism, are tedious to judge by eye. This article presents a convolutional neural network that sorts pairs of microscope images into four focus and astigmatism classes so the check can run without an operator. Each sample is two shots of the same spot: one in the base position and one with a shifted working distance, because the shifted shot is what makes astigmatism visible. The decisive ingredient is a custom Fourier-transform preprocessing layer that converts each input to its magnitude spectrum before the convolutional part sees it. On a dataset of about three thousand image pairs, the base model variants with-out the Fourier layer reached 49–65% test accuracy, while the corresponding variants with the Fourier layer reached 89–94%; the best result was 94% with channel-axis concatenation. The same shallow model beat three pretrained ImageNet networks (MobileNetV2 87%, EfficientNetV2L 52%, ResNet50 44%), which we did not expect. The best fusion of the two images was concatenation along the channel axis together with the Fourier layer.

POST-QUANTUM CRYPTOGRAPHY: CURRENT STATE OF STANDARDIZATION AND DEPLOYMENT

Tomáš Sochor, Hana Sochorová

DOI: 10.11118/ecos2026_034

The paper reviews the quantum computing advances and their impact on classical cryptography. the aspect of potential breaking of present (believably secure) cryptographic algorithms. Recent progress in the abilities of quantum computers as well as the standardization of new generation of cryptographic algorithms resistant to quantum-computing attacks is summarized. The classical cryptographic algorithms are reviewed and corresponding threats from quantum algorithms were summarized. In addition, existing symmetric and hash algorithms under quantum attacks is discussed and recommendations for improving security in long-term are formulated.

A UNIFIED FORMAT FOR AGGREGATING PRE-TRAINED POSE ESTIMATORS INTO CUSTOM HUMAN-SKELETON DATASETS

Damian Sova, Oldrich Trenz

DOI: 10.11118/ecos2026_035

Human pose estimation works well on benchmark datasets but loses accuracy once a model is deployed on real footage with varying distances, crowds and occlusion. Building a labelled dataset for each new deployment is expensive, so this work asks whether existing pre-trained detectors can be combined to generate such a dataset cheaply, without retraining any of them. The proposed Unified Format runs three public detectors (PoseNet, MoveNet, MMPose), maps their native keypoints into one common 27-keypoint skeleton, matches detected people across models and fuses them by weighted average. The method was evaluated on the COCO WholeBody validation subset (2,693 images with humans, 11,004 ground-truth instances) using OKS, average percentage error and mean squared error. The fused skeleton provides usable positional localisation (APE 3.3%, MSE 1,213.84) and predicts slightly more pose instances than MMPose alone (70.39% vs 69.84%), but it remains weaker than MMPose on the main accuracy-oriented metrics (OKS 0.23 vs 0.28; correct keypoints 37.93% vs 43.62%). The result is an honest, reproducible tool chain rather than a new state of the art.

TREESTORAGE: HIERARCHICAL STORAGE WITH VERSIONING AND INHERITED ACCESS CONTROL IN A UNIVERSITY INFORMATION SYSTEM

Petr Straka, Oldrich Trenz

DOI: 10.11118/ecos2026_036

Large information systems accumulate knowledge that is often transferred informally rather than stored in the system itself. This article describes the design and working prototype of Treestorage, a reusable hierarchical storage component for INET, the economic and personnel information system of Masaryk University. Treestorage is intended to organise agenda-related HTML notes, references to binary documents and parameterised SQL reports in a tree structure, with read and edit rights inherited from parent nodes and version history stored for edited nodes. Before implementing the component, available alternatives at the university were assessed, especially Google Drive, OneDrive and git-based services such as GitHub/GitLab; these options were not selected because they would either keep data and authentication outside INET or fail key requirements such as fine-grained folder-level read access. The implemented data model uses an adjacency list, reads nodes as a flat list and rebuilds the tree in the front end, while inherited rights are resolved through a PostgreSQL materialised view based on a recursive query. The result is a functional prototype that can be tested internally and further integrated after related INET 2.0 components are completed.

WEB APPLICATION FOR GENERATING TECHNICAL DOCUMENTS USING LARGE LANGUAGE MODELS

Martin Šulc, Ivo Pisařovic

DOI: 10.11118/ecos2026_037

This paper investigates the viability of leveraging large language models (LLMs) to generate formal corporate documents, moving beyond rigid, traditional template-based approaches. The study aims to verify whether combining reference samples, natural language instructions, and parallel multi-model processing can yield a practical enterprise tool. The proposed microservices architecture utilizes a Node.js/LangChain backend to abstract calls across four LLM providers (OpenAI, Anthropic, Gemini, and Ollama). This is supported by Python microservices for input parsing and multi-format document generation (Word, PDF, Excel), a Vue.js frontend wizard, and PostgreSQL/Redis persistence. System validation involved end-to-end scenarios across five document types, benchmarking seven models, and user testing with ten participants. Results demonstrate the solution’s efficacy. Furthermore, comparative analysis reveals that local open-source models achieve generation quality comparable to mid-range cloud services for Czech-language tasks, albeit with trade-offs in increased latency and initial hardware investments.

AI METHODS FOR STOCK-PRICE FORECASTING IN THE INDIAN MARKET: THE EFFECT OF MACROECONOMIC FACTORS

Ankit Tripathi, Oldřich Trenz

DOI: 10.11118/ecos2026_038

This article compares four artificial-intelligence forecasting methods (LSTM, SARIMAX, XGBoost and FB-Prophet) on a decade of daily prices of Reliance Industries, one of the largest companies on the Indian stock market. Most published forecasting studies use US, Chinese or European data; the Indian market is less frequently used as a test bed. In the source thesis, a Scopus query for “time series forecasting in India” refined by “financial market” returned 137 documents, while the broader bibliometric dataset contained 2,202 documents. The dataset covers 2 December 2013 to 30 November 2023 (over 2,600 trading days). Each model was first trained on price features alone and then re-evaluated after merging five macroeconomic indicators (gold price, USD/INR, the 10-year government-bond yield, the Wholesale Price Index and the Consumer Price Index). Accuracy was measured by MSE, MAE and the coefficient of determination R². Adding macroeconomic factors did not help every model. XGBoost improved across all three metrics (MSE 39.35 to 24.58) and was the only method that gained without producing implausibly perfect numbers; LSTM improved only with the larger 90% training split, while FB-Prophet overfitted badly (MSE 11.17 to 1,912.35). The results are specific to one company and a retrospective forecasting experiment on historical data; they do not represent live trading performance.

LARGE LANGUAGE MODEL AGENTS AS INVESTMENT ADVISORS: A BACKTESTED COMPARISON OF GPT-3.5, GPT-4 AND GPT-4-TURBO OVER INDIAN STOCK DATA

Ankit Tripathi, Michaela Veselá, Oldrich Trenz

DOI: 10.11118/ecos2026_039

This article asks whether an OpenAI GPT model can act as a conversational investment advisor when it reads structured stock data instead of free text. A pandas DataFrame of daily prices for Reliance Industries (RELIANCE.NS), spanning 1 April 2019 to 28 March 2024, was enriched with the signals of four hard-coded technical strategies: gap trading, trend following, Fibonacci retracement and mean reversion. Each strategy was backtested with Pyfolio, and three LangChain DataFrame agents wrapping GPT-3.5-Turbo, GPT-4 and GPT-4-Turbo answered the same seven plain-English questions. Trend following gave the highest backtest return (annual 13.60%, cumulative 86.60%), while Fibonacci retracement was the least risky (max drawdown only -1.10%, Omega ratio 6.33). The GPT-4-Turbo agent gave the most context-aware answers; GPT-3.5-Turbo was inconsistent and once misread a signal. The results hold for a single asset and a no-cost in-sample backtest, so they describe a feasibility study rather than live trading performance.

AN OFFLINE RAG SYSTEM PROTOTYPE FOR SAFE QUERYING OF UNSTRUCTURED DOCUMENTS

David Vlastník, Jaromír Landa

DOI: 10.11118/ecos2026-040

In the context of large language models (LLMs) being utilized within organizations that manage sensitive data, two notable limitations of commercial cloud solutions become evident. Firstly, there is the risk of confidential information being disclosed when content is transmitted to external servers. Secondly, there is the issue of the models‘ propensity to hallucinate in the absence of reliable context. The present paper proposes and implements a prototype of a fully offline RAG (Retrieval-Augmented Generation) system using exclusively open-source components. The architecture is predicated on the Agentic RAG approach, which substitutes linear query processing with a multi-agent system orchestrated using the LangGraph framework. A hybrid (dense + sparse) approach over the Qdrant vector database is used for search, and the parallel relational database DuckDB is used for tabular data. The system is containerized using Docker Compose and runs a local LLM via the Ollama platform. The evaluation of the output quality was conducted through a comparative analysis with three existing tools (LM Studio, AnythingLLM, NotebookLM) on a set of ten extensive PDF reports on climate change. The evaluation process involved the utilization of five complex questions, necessitating the synthesis of information from multiple sources. The implemented prototype achieved a level of accuracy comparable to the cloud-based NotebookLM tool and significantly outperformed solutions employing local LLMs. It also maintained transparency of answers through citations of source documents and ensured that data remained on the user‘s local device.

THE ROLE OF HAPTIC PRODUCT EXPERIENCE IN CONSUMER EVALUATION: AN EYE-TRACKING AND PUPILLOMETRY STUDY OF REUSABLE BOTTLES

Jan Vrána, Stanislav Mokrý, Michal Pšurný

DOI: 10.11118/ecos2026_041

This study examines how physical and digital product presentation shape consumer evaluation in the context of reusable water bottles. The research focuses on whether the possibility of direct tactile interaction with a product influences perceived attractiveness, purchase intention, certainty of evaluation, and pupil response. An experimental design combining eye-tracking, pupillometry, and a post-exposure questionnaire was used to compare electronic product presentation with physical presentation involving product handling. The findings suggest that tactile interaction did not systematically change perceived product attractiveness. However, for selected products, the form of presentation affected purchase intention and certainty of evaluation, indicating that haptic information may become relevant when material or functional product characteristics are salient. The pupillometry results showed differences between physical and electronic presentation, although pupil dilation was not consistently associated with product attractiveness or purchase intention. Overall, the study contributes to sensory marketing research by showing that physical product experience may influence selected aspects of consumer decision-making, while its effect depends on the specific product and evaluated attribute. The results also highlight both the potential and limitations of pupillometry as a complementary method for studying consumer responses to haptic product presentation.

FOREIGN DIRECT INVESTMENT, OIL REVENUES, AND TRADE AS DETERMINANTS OF ECONOMIC GROWTH IN GCC COUNTRIES

Evans Yeboah

DOI: 10.11118/ecos2026_042

This study investigates the roles of foreign direct investment (FDI), oil revenues, and trade in shaping economic growth in Gulf Cooperation Council (GCC) countries over the period 1995–2024. The analysis employs the pooled mean group autoregressive distributed lag (PMG-ARDL) model to capture both long-run and short-run dynamics, and uses fully modified ordinary least squares (FMOLS) for robustness. The results indicate that FDI and oil rents exert positive, statistically significant effects on long-run per capita income. In contrast, merchandise trade does not exhibit a significant contribution to growth. The short-run estimates confirm the presence of a stable adjustment mechanism toward the long-run equilibrium. These findings highlight the dominant role of resource revenues and external capital inflows in the GCC growth process and point to the need for policies that promote diversification and strengthen non-oil sectors.

FOMO EFFECT AND GENERATION Z’S SENSITIVITY TO MARKETING COMMUNICATION TOOLS IN THE CLOTHING INDUSTRY

Marek Záboj, Lenka Procházková, Karolína Malinová

DOI: 10.11118/ecos2026_043

The paper examines the influence of the FOMO effect (Fear of Missing Out) on the consumer behaviour of Generation Z in the clothing industry. The aim is to identify factors influencing purchasing decisions with an emphasis on the role of FOMO and to formulate recommendations for marketing communication differentiated according to consumer types. Empirical data come from a quantitative questionnaire survey conducted in October 2025 (n = 295) among respondents born in 1995–2010. Using hierarchical cluster analysis (Ward‘s method, Euclidean distance), respondents were divided into four segments according to the degree of FOMO experienced. Six hypotheses were tested using multiple linear regression (H1–H3) and Spearman‘s correlation coefficient (H4–H6). The results show that FOMO is a statistically significant predictor of sensitivity to time limited offers and sales promotions (p < 0.001) and sensitivity to brand communication content (p < 0.001). At the same time, a positive association was confirmed between the level of FOMO, the intensity of social media use and the frequency of following influencers. On the other hand, FOMO did not appear to be a predictor of the frequency of online purchases or the consideration of sustainability. The results suggest differentiated recommendations for brand communication, which for individuals with high FOMO emphasizes urgency and visual dynamism, while for segments with low FOMO, credibility, transparency and value content.

VISUALIZATION OF ALGORITHMS FOR GENERATING PSEUDORANDOM NUMBERS

Štěpán Zálešák, Radim Farana

DOI: 10.11118/ecos2026-044

This article deals with the design and implementation of an interactive web application intended for the visualization and comparison of two basic algorithms for generating pseudorandom numbers (PRNGs) — the Linear Congruential Generator (LCG) and the Linear Feedback Shift Register (LFSR). The work responds to the problem of the abstractness of teaching these algorithms and offers a tool that allows users to experiment with parameters, monitor the internal states of the generators, and analyze the quality of the resulting sequences using integrated analytical tools such as a histogram of value distribution, period detection, and a scatter plot to reveal the lattice structure.

Návrat zpět