Public Health Institute of the Republic of Srpska

Political decisions made under geopolitical conditions significantly influence global digital‎ ecosystems, as illustrated by the U.S. ban on TikTok and Germany’s withdrawal of advertising from the‎ platform X (formerly Twitter). This study identifies and predicts patterns of digital platform fragmentation ‎emerging from such politically driven interventions. By integrating social network analysis, sentiment ‎analysis, and predictive modelling, the research connects regulatory behaviour, public sentiment,‎ and predictive scenarios to demonstrate how political and corporate strategies jointly reshape global‎ communication structures. The findings reveal deep polarisation of public opinion on digital regulation, ‎while predictive results suggest that ongoing fragmentation may accelerate the formation of regional‎ platforms and increase digital isolation. The paper concludes with recommendations for policymakers to‎ develop adaptable legal and governance frameworks capable of addressing the geopolitical complexity ‎of contemporary digital environments.‎

The paper presents a chance find, designated as a grave assemblage from the site of Dobor near Modriča in northern Bosnia. The assemblage consists of three bronze bracelets, a bronze pendant, eleven bronze calotte-shaped buttons, and a fragmented ceramic vessel. The bronze pendant of the Ghidici type provides a basis for dating the deposited objects to the first half or middle of the eighth century BC at the latest, i.e. to the Ha C1 of the Central-European chronology. The micro-location of the Early Iron Age burial is directly linked with the settlement site of Doborski Brijeg and the excavated Late Bronze and Early Iron Age layers. The fragments of pottery decorated in the Basarabi style suggest the regional importance of the settlement as a possible intermediary in the exchange of cultural goods and influences coming from neighbouring regions, primarily Slavonia and Syrmia (Srem/Srijem), along the Bosna river valley.

In recent years, layered double hydroxide (LDH) has developed strongly in the field of corrosion protection of materials due to its special properties, which include exceptional anion exchange capability with increased anion capacitance as well as barrier resistance and a recorded structural memory effect. This article reviews recent work on the properties of LDH in the form of powders and films. The possibilities and advantages of using this compound to protect materials from corrosion are analyzed in detail.

Nature Park "Orjen", located in the far south of the Republic of Srpska, represents one of the richest floristic areas in Bosnia and Herzegovina. The aim of this research is to systematically show the presence of endemic and subendemic vascular plants in this area, and to analyze their taxonomic and ecological structure, as well as their protection status. The research recorded 159 taxa, of which 118 belong to Balkan endemic plants, and 41 to subendemic plants. The analysis of life forms shows the dominance of hemicryptophytes (52.2%) and chamaephytes (27%), which is in accordance with the general characteristics of the Balkan endemic flora of BiH. Taxonomically, the most represented families are Asteraceae, Caryophyllaceae, Fabaceae, Lamiaceae and Poaceae. The special conservation value of the park is confirmed by the fact that as many as 123 recorded taxa have the status of strictly protected or protected species according to the Regulation on Strictly Protected and Protected Wild Species of the Republic of Srpska. These results emphasize the exceptional importance of the Nature Park "Orjen" for the preservation of biological diversity and represent the basis for future activities in the field of nature protection, as well as for further floristic and ecological research.

Postal traffic and transport face challenges related to the rapid growth of parcel volumes, increasing demands for sustainability, and the need for integration into the smart transportation concept. This study explores the application of machine learning (ML) models for the classification of postal delivery times, with the aim of improving service efficiency and quality. As a case study, the Postal Center Zenica, one of the seven organizational units of the Public Enterprise “BH Pošta” in Bosnia and Herzegovina, was analyzed. The available dataset comprised 11,138 instances, which were cleaned and filtered, then expanded through two iterations of data augmentation using an autoencoder neural network. Five ML models, Random Forest, Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), k-Nearest Neighbors (kNN), and Multi-Layer Perceptron (MLP), were developed and compared, with hyperparameters optimized using the Bayesian method and evaluated through standard classification metrics. The results indicate that the data augmentation method significantly improves model performance, particularly in the classification of delayed shipments, with ensemble, especially Random Forest and XGBoost, emerging as the most robust solutions. Beyond contributions in the context of postal traffic and transport, the proposed methodological framework demonstrates interdisciplinary relevance, as it can also be applied in telecommunication traffic classes, where similar network dynamics require reliable predictive models.

For upgrade of public transport importance of demanded service quality from consumers is of the greatest importance for many reasons, like budget and technical limitations, external and internal demands, projected quality of services and delivered services. Purpose of this research is description and analysis of difficulties in providing public transport on Doboj city territory. Aim of this report is explore provided services of existing public transportation system, detect advantages and flaws, and propose actions for it’s upgrading. SWOT analysis will provide information for current benefit or disadvantage, and show opportunities and risks for system’s future also assessment of possible corrections in functionality.

BackgroundMonitoring food marketing that reaches children is key to effective implementation of any regulation. The Best-ReMaP European Union Joint Action aimed to test the feasibility of using monitoring protocols (developed by the World Health Organization’s Regional Office for Europe [WHO-Euro]) in the public health sector, to inform EU-wide monitoring plans. In Republic of Srpska, BiH, we piloted the WHO ‘CLICK’ Framework step, ‘L – Landscape of campaigns’, created to assess the nature of digital food and beverage marketing that children are potentially exposed to when using their preferred digital media.MethodsThis mixed methods pilot study analysed 235 Instagram and Facebook posts from 10 brands/products popular among children in Republic of Srpska, and 24 videos of two YouTube influencers, between January-April 2023. The WHO Nutrient Profile Model (1st edition, 2015) was used to assess ads’ eligibility to be marketed to children. With WHO-Euro monitoring Protocols we assessed the promotional techniques (ad’s creative power) used by brands and YouTube influencers. Finally, we assessed feasibility at every step of the process.ResultsResearch capacity and coding were among the main identified challenges in the pilot. Our learnings contributed to updates of the WHO Protocols and the EU- framework guidance. Regarding the food marketing landscape, the most frequently advertised food categories on social media and in YouTube Influencers’ content were cakes, sweet biscuits and pastries (31% and 22.7% respectively) and chocolate and sugar confectionery (27.5% and 14.7% respectively). None of the social media ads were eligible to be marketed to children and only of 2 of the 24 YouTube influencers’ marketing instances were eligible. Most ads featured elements potentially appealing to teens, 88% on Facebook and 72% on Instagram. The most frequently used persuasive appeals on social media were reference to holidays, travel, or adventure (23%), followed by taste (12.5%) and premium/contest (11.5%). YouTube influencers predominantly presented branded foods (62.5%). Only one influencer video featured a marketing disclosure. Food cue contexts varied between the two influencers, depending on their target groups. One mostly posted videos filmed at home, with child appeals and referred positively to the food featured; the other presented food as a secondary object and in neutral rather than positive contexts.ConclusionThe WHO-Euro ‘landscape’ Protocol is a feasible tool for Member States to evaluate digital food marketing. Facilitators were: capacity/resources, access to a knowledge/experience exchange, and regular protocol updates. The analysis shows that in Republic of Srpska, BiH, powerful marketing of unhealthy foods appealing to children takes place on children’s favoured social media platforms and via influencers with child-appealing content. Mandatory, government-led restrictions, accompanied by regular monitoring are needed to protect children’s health.Supplementary InformationThe online version contains supplementary material available at 10.1186/s41043-025-01009-8.

Abstract Consumers of agricultural products represent a sensitive category of society, therefore communication with them is particularly important. Communication is influenced by a large number of factors, and in this regard, making the right decision about the way to communicate is an extremely complex job. In previous research, the emphasis was placed mostly on the entire supply chain, much more than on the method of choosing communication with end consumers of agricultural products. Therefore, this research represents an attempt to make an additional contribution to rational decision-making in this part of the agricultural supply chain. In this paper, the selection of criteria used in the process of communicating with consumers in a country was carried out by applying expert fuzzy multi-criteria decision-making. The innovative fuzzy method SiWeC (Simple Weight Calculation) was applied to evaluate ten given criteria with the aim of their selection. The subject of the research were several agricultural companies and the experts were six experts from the marketing department of the companies in question. The results show that the method of communication itself is the best evaluated criterion, as well as the delivery time, i.e. the impact of delivery speed on the demand for a particular agricultural product. Uncertainty in decision-making is reduced by applying fuzzy decision-making logic, and the results provide the basis for future research that should be directed at selecting the most favorable methods of communication when products of this type are in question, as well as the development of the existing methodology, which in this case has proven to be adequate.

The experiments were conducted on the demonstration field of the Agricultural Institute of the Republic of Srpska, from 2021 to 2023. The aim of this research was to examine the effect of applied fertilization treatments on the yield of grain and harvesting residues of broad beans under the agro-ecological conditions of the Banja Luka region. Sowing was done manually in the second half of March each year of the study. The experiment was set up in four replications using a randomized block design, with the size of the basic plot being 10 m². For these trials, seed from the broad bean population from the Institute's collection was used, with a sowing rate of 220 kg ha⁻¹. During the study, four fertilization treatments were tested: control (T1), NPK 15:15:15 at a dose of 300 kg ha⁻¹ (T2), biostimulator Slavol S at a dose of 250 ml ha⁻¹ (T3), and NPK 15:15:15 at a dose of 300 kg ha⁻¹ + biostimulator Slavol S at a dose of 250 ml ha⁻¹ (T4). The following parameters were monitored during the study: grain yield per unit area (kg ha⁻¹), harvesting residue yield (kg ha⁻¹), biological yield (kg ha⁻¹), the share of empty dry pods in the harvesting residues (%), the share of dry matter in the plant in the harvesting residues (%), and the harvest index (HI, %). During these experiments, the grain yield of beans ranged from 1,096.7 kg ha⁻¹ in treatment T1 (2021) to 3,240.0 kg ha⁻¹ in treatment T4 (2023). The highest average grain yield during these studies was achieved with treatment T4, amounting to 2,679.9 kg ha⁻¹. The average yield of harvesting residues during the study ranged from 3,049.7 kg ha⁻¹ (T1) to 3,986.8 kg ha⁻¹ (T4). The share of empty dry pods in the total harvesting residue yield varied from 25.5% in treatment T1 (2022) to 36.5% in treatment T4 (2021). On the plots where only the biostimulator (T3) was used, the lowest harvest index was recorded (31.5%), while the highest was achieved with treatment T2 (40.9%).