Solid-phase microextraction (SPME) has emerged as a powerful tool used for extracting analytes from complex matrices, offering high sensitivity, automation potential, and high-throughput efficiency. Despite its advantages, its performance is often limited by the selectivity and physicochemical properties of conventional coatings. This study addresses the limitations of conventional SPME coatings for monitoring environmental contaminants by introducing a nitrogen-rich covalent organic framework (COF) as an SPME extraction phase/coating. The COF-based SPME coating was implemented in a coated-blade spray (CBS) device format and was used to develop a robust analytical workflow that combined ambient mass spectrometry ionization via CBS for rapid screening, followed by LC-MS/MS for confirmation and quantification. Statistical validation using a t test showed that only 3 of 43 analytes exhibited statistically significant differences in loss during the microdesorption step of the CBS-MS/MS screening. The results demonstrate that the COF-based SPME coating facilitates simultaneous extraction of analytes with varying polarities from river samples without altering the sample matrix. The final protocol achieved limits of quantification ranging from 0.025 to 5 ng mL–1. The method was successfully applied to river samples, quantifying the target analytes and demonstrating that LC separation remains essential for confirmation. This work highlighted the effectiveness of a COF-based SPME coating as a next-generation extraction phase that simplifies rapid ambient MS screening, followed by traditional LC-MS/MS for confirmation and quantification, all within a single coated SPME device.

We study the phenomenology of the topological portal between QCD and pseudo-Nambu--Goldstone dark matter proposed in~\cite{Davighi:2024zip}. We construct a weakly coupled ultraviolet completion in which a vector mediator gauges a baryonic current of QCD for the light quark flavours, linking it to dark pions parametrizing the coset $SU(2)_D/U(1)_D$. As a concrete anomaly-free realisation, we gauge the leptophobic combination $B_1+B_2-2B_3$. This completion allows us to calculate thermal coannihilation beyond the regime of validity of the pion effective theory, including resonant mediator effects, and to test the resulting thermal target with boosted-dijet and monojet searches at the LHC, with LEP and bottomonium data, and with flavour observables. These collider signatures follow directly from the symmetry structure required by matching onto the topological portal. At low momentum transfer, the topological operator instead controls the decay of the heavier dark pion, which is long-lived. A future high-statistics $Z$ factory, such as FCC-ee, can probe the low-mass region where hadron-collider sensitivity deteriorates, providing complementary tests of the topological portal.

This study examined the association between students' classroom use of general ICT applications and computer and information literacy (CIL) across 35 education systems participating in ICILS 2023. Jurisdiction-specific weighted regression models were estimated using all five CIL plausible values and 75 jackknife replicate weights while adjusting for socioeconomic background, sex, and age. Jurisdiction-specific coefficients were synthesized using random-effects meta-analysis. Linear models indicated a modest positive association between classroom ICT use and CIL, but quadratic models consistently provided a better fit. The quadratic coefficient was negative in 34 of the 35 jurisdictions, yielding a pooled estimate of −5.66 (95% CI −6.58 to −4.74). Predicted CIL increased with classroom ICT use to approximately the midpoint of the international scale, after which gains diminished and CIL gradually declined. This nonlinear pattern was remarkably consistent across education systems and did not vary meaningfully by socioeconomic background. These findings suggest that the educational value of classroom ICT use depends not only on how often digital technologies are used, but also on how they are integrated into teaching and learning.

While the GDPR and ePrivacy Directive mandate that consent information must be clear and accessible, most modern cookie banners remain obscured by technical jargon, vague phrasing, and frequent content overload or underload. This feasibility study investigates the impact of applying plain language (Einfache Sprache) to cookie banners within the IAB Transparency&Consent Framework (TCF). In our study, we analysed cookie banner texts from 200 websites, using AI-based mapping to categorise extracted content into standardised processing purposes. By substituting complex legal terms with simplified descriptions, we successfully demonstrated that the comprehension barrier can be lowered from a college-graduate level to a 7th-grade level. However, the effectiveness of plain language is inherently constrained by the informativeness of the original content; it cannot compensate for banners that omit legally required details. We conclude that while plain language is a vital tool for digital accessibility, it must be paired with standardised implementation guidelines to ensure that cookie banners are both readable and informative.

Physics-informed anomaly detection for industrial control systems (ICSs) is usually validated on synthetic data. How much of that performance survives on real hardware-in-the-loop (HIL) data is an open question. We re-implement PhySec-Edge, a hybrid BiLSTM/β-VAE/GAT/Kalman pipeline, as a full architecture and evaluate it on HAI 23.05. The gap is consistent across five seeds. AI-only F1 falls from 0.774 (synthetic) to 0.525 (HAI); the full hybrid falls from 0.798 to 0.512. The components diverge. PVE residual sharing transfers but weakens on real data (ΔF1 = +0.010 ± 0.008 across five seeds against +0.017 synthetic; ΔMCC = +0.013, positive in every seed). The PVE NIS decision score fails outright (raw FPR = 0.482; a correction brings it to 0.013, but detection does not recover). Under the retrospective fixed-FPR comparison, the preferred configuration is GNN+BiLSTM (F1 = 0.530 ± 0.002), outperforming full-EADE fusion (0.512 ± 0.003). From this split we derive the information-vs-decision pathway distinction, a way to reason about component-level generalization in hybrid architectures. We then set out what it means for physics-informed ICS detector design and edge deployment.

This paper presents a dual-layered AI framework for real-time landslide risk assessment developed under the GeoNetSee project within the Interreg Danube Region Programme. The first layer employs a fuzzy logic model, inspired by the Slovenian MASPREM system, which integrates Landslide Susceptibility Maps (LSS) with high-resolution precipitation forecasts from the Open-Meteo API to generate a Predicted Landslide Hazard (PLSH) score on a 0–5 scale, updated every 6–12 hours. The second layer focuses on real-time ground displacement detection by fusing low-cost dual-frequency GNSS receivers with MEMS accelerometers and applying machine learning regression algorithms, including Support Vector Regression (SVR), Long Short-Term Memory (LSTM), and Convolutional Neural Network – Long Short-Term Memory (CNN-LSTM). Validation at the Curine Njive pilot site in Bosnia and Herzegovina demonstrated that SVR achieved a Mean Absolute Error (MAE) as low as 0.0018 m across four GNSS nodes, while the CNN-LSTM achieved the highest overall predictive accuracy across all evaluated nodes with a MAPE of only 3.99% for Node 41. The fuzzy model successfully identified critical hazard levels during intense rainfall events. Results confirm the viability of integrating low-cost sensor networks with AI-driven analytics for transboundary geohazard early warning systems.

Fuzzy inference systems have demonstrated considerable promise for EEG-based cognitive state monitoring in neurodegenerative conditions. However, two design decisions significantly influence system performance and clinical applicability: the choice of inference architecture (Takagi-Sugeno vs Mamdani) and the method of rule and membership function generation (manual expert-driven vs data-driven automated). This paper presents a comparative analysis of both dimensions in the context of an EEG-based Alzheimer’s disease monitoring system operating on the ds004504 OpenNeuro dataset (88 subjects: 36 AD, 23 FTD, 29 HC). A Takagi-Sugeno system, implemented as a hybrid FSM-Fuzzy architecture, is compared against a Mamdani equivalent across four axes: inference transparency, output interpretability, computational complexity, and suitability for FPGA deployment. The limitations of manual membership function tuning are then examined through failure case analysis, and an Adaptive Neuro-Fuzzy Inference System (ANFIS) proof-of-concept is presented, demonstrating automated rule and MF generation from the delta/alpha ratio (DAR) biomarker. ANFIS training on 88-subject DAR data (70 training, 18 test) achieved minimal training RMSE of 0.392 and validation RMSE of 0.391. Data-driven MF centers differ substantially from manually designed centers — shifts of 2.754σ, 0.455σ, and 2.055σ for LOW, MEDIUM, and HIGH MFs respectively — quantifying the representational gap that expert tuning introduces. These findings motivate a hybrid design strategy: ANFIS for offline MF calibration, TS for online FPGA inference, and Mamdani for clinical output presentation.

This paper evaluates the technical, economic, and environmental feasibility of residential prosumer implementation in Sarajevo, Bosnia and Herzegovina, through the integration of photovoltaic (PV) generation and battery energy storage. A simulation‑based assessment was performed in HOMER Pro using a measured 12‑month household load profile with an average daily consumption of 11.25 kWh and a peak demand of 2.85 kW. Six system configurations were investigated, consisting of 3 kW, 4 kW, and 5 kW PV installations, each analyzed with and without a 5.04 kWh lithium‑ion battery, over a 25‑year project lifetime under a zero‑credit export scheme. Among the analyzed configurations, the 3 kW PV system without battery storage achieved the lowest Net Present Cost (NPC) and Levelized Cost of Energy (LCOE) of 0.137 BAM/kWh, representing a 28% reduction compared to the grid tariff, with a simple payback period of 17 years and the highest Internal Rate of Return (IRR) of 3.8%. The 4 kW PV system provided the most balanced operational performance, generating 4,954 kWh annually with imports and exports nearly equal, yielding an LCOE of 0.139 BAM/kWh and a 20‑year payback. The 5 kW PV system generates highest total renewable energy and full annual carbon neutrality, but its surplus generation provided limited additional financial benefit under the current net‑metering framework, which only offsets future bills rather than generating revenue. Battery storage integration substantially enhanced energy autonomy by increasing self‑consumption and renewable fraction (from 52% to 61% for the 3 kW system, and from 57% to 68% for the 4 kW system), but it raised LCOE to 0.198–0.202 BAM/kWh due to higher capital investment. The environmental assessment demonstrated that PV adoption significantly reduces greenhouse gas emissions, with the 3 kW system lowering annual CO₂ emissions by ~76%, the 4 kW system by ~92%, and the 5 kW system achieving net‑negative emissions. The results demonstrate that under zero‑sellback regulatory frameworks, PV sizing must prioritize direct self‑consumption matching. Moderate‑capacity PV installations therefore represent the most economically viable investments, while battery storage serves primarily as a driver of energy independence rather than immediate financial gain.

The primary objective of this research is to examine the direction and strength of the impact that capital structure indicators have on the performance of banks in Bosnia and Herzegovina. Specifically, the study aims to assess the degree and nature of the interdependence between capital structure and profitability indicators within the banking sector of Bosnia and Herzegovina. The empirical analysis is based on audited annual data from commercial banks operating in the Bosnian market from 2008 to 2024. To explore these relationships, econometric panel models with fixed and random effects were employed. The findings reveal a statistically significant influence of the debt-to- equity and debt-to-asset ratios on bank profitability. Additionally, internal factors such as bank size and the cost-to-income ratio demonstrated notable significance. Beyond capital structure variables, the analysis also incorporated selected macroeconomic factors, which were found to play an important role in shaping bank profitability. The results highlight the relevance of the regulatory framework and prevailing market conditions in Bosnia and Herzegovina, as they considerably affect bank operations and their ability to generate profit.

This paper proposes an integrated Expert Criteria Selection and Weighting (ECSW) model for expert-based criteria evaluation, selection, and weighting in multi-criteria decision-making (MCDM). The model is designed to preserve the expert-driven nature of the decision process while reducing the cognitive burden and potential inconsistency associated with pairwise comparisons and criteria ranking. It integrates four components: statistical criteria selection, assessment of the informativeness of expert ratings, determination of criteria weights based on their relative importance and rating stability, and sensitivity analysis. Candidate criteria with below-average ratings are evaluated by testing whether their expert-specific deviations from the experts’ average rating patterns are significantly below zero, using a one-sided one-sample t-test followed by the Holm–Bonferroni correction for multiple comparisons. The final weights are obtained by adjusting equal base weights according to the relative importance of each criterion and the dispersion of expert ratings. Model stability is examined using a leave-one-expert-out procedure and Pearson and Spearman correlation coefficients. The weighting procedure was preliminarily verified against results obtained using the Ordinal Priority Approach. A strong agreement was observed for both criteria weights (r = 0.956) and rankings (ρ = 0.929), with p < 0.001 in both cases. The two models also identified the same ten highest-ranked criteria, although some differences occurred in their ordering. The ECSW model is particularly suitable for decision problems involving a large number of criteria for which conventional pairwise-comparison or ranking-based procedures may be difficult to apply. A supplementary ECSW implementation tool is provided to facilitate practical application and reproducibility.

Natural zeolites may exhibit certain limitations in the adsorption of metal ions from aqueous solutions due to their heterogeneous mineral composition, variable content of exchangeable extra-framework cations, and the presence of impurities. In this study, natural clinoptilolite zeolite was modified using an ethanolic extract derived from lemon peel waste to evaluate the effect of this sustainable modification approach on Cu(II) adsorption from synthetic aqueous solutions. Batch adsorption experiments were conducted using 50 mL of Cu(II) solution at 300 rpm for 60 min. A one-factor-at-a-time (OFAT) approach was applied to investigate the effects of initial solution pH (2–6), initial Cu(II) concentration (10–100 mg/L), and adsorbent dosage (500-2500 mg) for both unmodified natural zeolite (UN-zeolite) and fruit-waste-extract-modified zeolite (FWEM-zeolite). All experiments were performed in triplicate. At pH 6, complete Cu(II) removal (100%) was achieved using UN-zeolite, whereas FWEM-zeolite reached a removal efficiency of 84.37%. At initial Cu(II) concentrations of 10 and 50 mg/L, complete removal was achieved using UN-zeolite, whereas FWEM-zeolite achieved removal efficiencies of 87.97% and 80.91%, respectively. Increasing the UN-zeolite dosage resulted in complete Cu(II) removal at 2000-2500 mg, while FWEM-zeolite reached its maximum removal efficiency of 91.9% at a dosage of 1000. Overall, modification with lemon peel waste extract did not improve Cu(II) removal compared with the unmodified natural zeolite under the investigated conditions. Nevertheless, FWEM-zeolite maintained high Cu(II) removal efficiency, particularly at elevated initial concentrations, demonstrating the potential of fruit-waste-derived extracts as sustainable modifying agents for natural mineral adsorbents

Multi-criteria decision-making methods conventionally rank alternatives by aggregating normalized criterion values into a composite score, an approach that depends on the normalization scheme, requires externally determined criterion weights, and evaluates performance in absolute terms without accounting for inter-criteria dependencies. This paper proposes machine learning-based Peer-Prediction Trees for multi-criteria decision-making, a novel method that ranks alternatives by their excess performance over data-driven peer expectations. For each criterion, a leave-one-out cross-validated decision tree predicts each alternative’s value from its performance on all other criteria; the residual between observed and predicted values extracts how much the alternative outperforms the empirical trade-off structure constraining its peers. Residuals are standardized, aligned with the preferred criterion direction, and aggregated into an Excess Performance Index using uniqueness-based weights derived endogenously from the cross-validated predictive fit. The method is validated on three real-world datasets: a numeric vehicle selection problem (22 alternatives, five criteria), a fully linguistic financial performance evaluation of Turkish listed firms (32 alternatives, four criteria), and a mixed numeric-linguistic exchange-traded fund selection problem (12 alternatives, nine criteria). The contributions of this work are fourfold: (1) the introduction of peer prediction via leave-one-out cross-validated decision trees as a multi-criteria decision-making ranking paradigm that evaluates excess performance rather than absolute scores; (2) a uniqueness-based criterion weighting scheme embedded into the method, eliminating the need for separate weighting procedures; (3) native handling of numeric, linguistic, and mixed decision matrices within a single unified framework; and (4) empirical validation across three heterogeneous real-world applications demonstrating the method’s versatility and interpretability.