Abstract: Digitalization has brought IoT and big beehive data in beekeeping. Consequently, the need for advanced analysis of beehive data and other bee-related data has led to machine learning. The objective of the paper is to reveal research tendencies, trends and gaps in recent literature on machine learning advances in beekeeping. Therefore, the paper reviews articles that have been published in the last two years in good-quality journals. More specifically, the review analyzes and discusses applications, (sub)types and algorithms of machine learning, including data sets used, that have been reported in selected articles. The results have shown that the most common application cases are related to bee products quality assessment/authentication and identification/prediction of beehive conditions. To this end, supervised learning, more specifically, classification, has been predominantly used. The most often used algorithms are Random Forest, Support Vector Machine and Convolutional Neural Network. Moreover, a variety of data has served as input to the models including images, sound, data from beehive sensors, meteorological data, spectroscopy data on honey samples, etc. Machine learning in beekeeping is still in its early years. However, many opportunities have already been identified and promising research directions are just opening out.

Abstract: In this paper, we present an efficient classifier that is able to perform automatic filtering and detection of tweets with clear negative sentiment towards COVID-19vaccination process. We used a transformer-based architecture in order to build the classifier. A pre-trained transformer encoder that is trained in ELECTRA fashion, BERTic, was selected and fine-tuned on a dataset we collected and manually annotated. Such an automatic filtering and detection algorithm is of utmost importance in order to explore the reasons behind the negative sentiment of Twitter users towards a particular topic and develop a communication strategy to educate them and provide them with accurate information regarding their specific beliefs that have been identified.

Abstract: According to various reports, a significant amount of energy use and carbon emissions can be attributed to the building sector. Apart from increasing the utilization of renewable sources, demand flexibility has also been noted as a key factor that can positively contribute to the reduction of the carbon footprint of buildings. A key driver of demand flexibility are novel approaches in the thermal domain. Here, heat pumps present a noteworthy untapped potential with projections of even higher impact in the future. However, current solutions for heat pump simulation lack interoperability and ease of deployment. Hence, this paper presents an improved, open approach to both heat pump modeling as well as building modeling, such that the complete loop from controls inputs to indoor ambient temperature can be simulated.

Abstract: With the proliferation of heat pump systems for both heating and cooling applications for a wide range of space volumes, from isolated rooms to whole houses and buildings, their efficient operation is paramount to facilitate the transition to a more efficient building stock and reduction of greenhouse gas emissions. Also, phasing out polluting non-renewable fossil fuel-based heating systems in favor of heat pumps contributes notably to the electrification of the thermal domain and allows for a more notable share to be facilitated by clean and renewable generation in the future. Therefore, on top of modeling approaches for these types of systems, adequate control algorithms need to be developed and deployed to ensure the proper utilization of flexibility that these devices offer. This paper presents a set of techniques based on reinforcement learning for heat pump control of room temperature based on varying source and user loop flow rates as control inputs and discusses the implications of a selection of different control strategies on the observed indoor temperature variables.

Abstract: Introducing an additional dimension for system scaling can enhance the performances of the systems based on Blockchain technologies. The BloHeS (Blockchain Healthcare System) implements clustering and hierarchical consensus mechanism to support sufficient capacity for a global Blockchain-based healthcare system. The validators in the BloHeS system perform self-organizing procedures for clustering. They implement the Island management protocols for Node addition, Island splitting and Island merging. This paper presents the Island management protocols that are the first self-organizing clustering protocols in the domain of Blockchain technologies.

Abstract: It is safe to say that the constant rise of internet-based media has changed the way people learn about the modern world, important events, and personalities. The ease of accessing any news there is and the speed of it being spread over social media by the users themselves have been the most important factors in shifting towards a decentralized and democratic way of gathering information. However, even in that setting, there exist more and less influential sources and spreaders of information, that play a role in forming the public opinion about given topics. Can we identify who these sources are in today’s social media? On the receiving side, how prone are users to changing their minds when exposed to new information and to what extent? How does the structure of the social network around the users influence their opinion or change of mind? In this paper, we try to answer these questions through the example of, perhaps the best tennis player of all time, Novak Djokovic, and the recent drama ´ around him tied to his participation in the Australian Open 2022 tournament. Throughout our work, we use Twitter as the main source of data, utilizing the information carried in the tweets themselves, as well as the topology of the network of its users. For a given period of time, we analyze the sentiment of users’ tweets about Djokovic´ (whether it is positive, negative, or neutral), as well as the influence of popular Twitter users on their followers, i.e. how does the sentiment of their tweets spread across their followers. The results of our analysis are then compared between two disjoint time spans - after Djokovic won his record-tying 20th Grand Slam title at Wimbledon 2021 and during the aforementioned drama in January 2022. Finally, this comparison lets us inspect the trend of users potentially changing their minds about Djokovic after this debatable event.

Abstract: A development of contemporary industrial production systems is heavily influenced by the fourth industrial revolution. These systems are undergoing a shift from mass production to mass customization. To support mass customization, Industrial Control Systems (ICSs) should be implemented in a modular and flexible way. One way to achieve this goal is to apply a microservice architecture in the context of ICSs. Since such an architecture has not been frequently applied in ICSs recently, it is still considered as an open research challenge. The adoption of microservices in ICSs increases the number of Application Programming Interfaces (APIs) which could be burdensome both for developers who maintain API integrations and users who need to be aware of too many service endpoints. Therefore, a microservice architecture should include an API gateway module to alleviate these issues and provide a single point of connectivity. In this paper, we identify and describe the requirements for an API gateway solution to be applied in the domain of ICSs. These requirements are then used to identify one or more API gateways that are the most suitable to be applied in embedded systems .

Abstract: Biogas produced through the process of anaerobic digestion can utilise a wide range of different feedstocks and is considered a renewable energy source. The EU Commission has recognised the role of AD in achieving circular economy goals and set the biogas and digestate production as recycling in the hierarchy of waste-to-energy operations. This work presents a Geographic Information System (GIS) based approach used for mapping of the biogas potential, derived from manure generated in livestock farms. In accordance with the calculated and mapped potential, optimal locations of biogas plants were determined, and the shortest transport routes were determined by using the GIS tool. The presented method was tested at the case study of five Croatian counties.

Abstract: In this paper, we will analyze how we can use blockchain, biometric security, and virtual worlds to create a secure environment for immersive interaction. The current model, is either not immersive, does not have digital avatars, or the system is not protected in a secure way to ensure full trust between participants We describe an immersive system that is fully secured using blockchain and biometric security to ensure users are the owners and controllers of their data and that they are fully verified on the system. Finally, we will analyze how the process is being performed and are some of the use cases of that kind of immersive environment that industries/participants can benefit from.

Abstract: The need for blockchain technology is growing constantly and its advantages are applied in different fields. One of the most sophisticated applications is the electronic voting system. Building an e-voting system based on blockchain gives a decentralized system. This research gives an overview which attacks can happen to such a system which is based on blockchain, keeping in mind that this system exists. It appears that attacks based on mathematical fields can happen such as: correlational, algebraic and birthday attacks. We analyse their performances and give an overview how each attack can happen and how to avoid it from happening. When choosing the parameters for the electronic voting system we must be careful of the presence of various attacks. This analysis showed that all processed attacks that have a mathematical basis can occur under different circumstances. They may violate voter anonymity, as in the case of correlation and birthday attacks, or they may misuse voting on behalf of others, as in the case of algebraic attacks. Fast responses, observation, correct cryptanalysis and selection of well- examined parameters are the main factors that can contribute to reliable communication. Keywords blockchain technology, cryptanalysis, birthday attack, correlational attack, algebraic attack.

Abstract: In recent years, blockchain technology and smart contracts gained popularity in many industry fields. Industry leaders are interested in private blockchains that are more suitable for enterprise applications. An application of blockchain brings the unprecedented capability to monitor and execute collaborative business processes (CBPs) in finance, trade, insurance, healthcare, and other fields. CBP activities can initiate smart contracts’ actions. That results in creating transactions and storing data objects of interest for a particular business process. Organizations involved in blockchain networks can benefit from analyzing transaction data in terms of improving blockchain systems and increasing knowledge about CBPs. However, blockchain platforms store transactions in data formats that are highly optimized for blockchain operations. Such formats are not suitable for ad hoc querying or data analysis. Therefore, organizations need a proper way to query and analyze data generated within blockchains. In this paper, we propose HyperETL, an ETL-based approach for accessing, extracting, and transforming blockchain data into a platform-independent format suitable for data analysis. We focus on private blockchain frameworks from the Hyperledger foundation. We also present a prototype solution that follows the proposed approach. Such a solution would help data analysts query data and derive valuable insights on CBP activities in order to analyze CBPs and improve decision-making. Furthermore, extracted and transformed blockchain data can be integrated into existing data warehouse systems.

Abstract: Fierce competition and rapidly changing market conditions make it harder for Small and Medium-sized Enterprises (SMEs) to achieve business success. To deal with rising challenges, SMEs form Virtual Organizations (VOs) and seize business opportunities jointly. In this paper, we present an expansion of a novel methodological approach based on a Distributed Ledger Technology (DLT) that promotes trustworthy collaborative production execution within a non-hierarchical VO. The expansion utilizes collaborative production process models, designed using the CE-MultiProLan Domain-Specific Modeling Language (DSML), to automatically generate artifacts used for expediting the configuration of DLT networks. The automatic generation of DLT artifacts reduces the problem of the slow and manual process of setting up a DLT network for a VO and enables VOs to start their production promptly and achieve their deadlines earlier.

Abstract: The COVID19 pandemic has brought health problems that concern individuals, the state, and the whole world. The information available on social networks, which were used more frequently and intensively during the pandemic than before, may contain hidden knowledge that can help to better address some problems and apply protective measures more adequately. Since the messages on Twitter are specific in their length, informal style, figurative speech, and frequent use of slang, this analysis requires the application of slightly different techniques than those classically applied to long, formal documents. To determine which topics appear in tweets related to vaccination, we apply state-of-the-art topic modeling techniques to determine which one is the most appropriate. This kind of research is meant to give us an insight into the opinions of the Twitter community on the phenomenon of vaccination and all related aspects. Comparing the results of the LDA with the topics obtained by manual annotation over the same set, we concluded that the LDA method provides a very good interpretation of the topics. Such data allow the analysis of sentiment, in this case pro- or anti-vaccination attitudes, and of specific groups of data and topics.

Abstract: The recent trends and challenges emphasize the need to exploit the potential of disruptive technologies in public administration. Accordingly, the main aim of the paper is to examine this issue over the last two decades. The results of bibliometric analysis on 3595 documents from Scopus reveal the growth of disruptive technologies research in public administration over time, especially in the last decade, as accelerated by several of the most relevant documents published in reputable journals such as Government Information Quarterly, Sustainable Cities and Society and Sustainability by several prominent authors. Most research has been conducted in the United States, followed by the United Kingdom and China and focused especially on artificial intelligence, followed by the internet of things, social media and blockchain, with the smart city being an important concept in disruptive technologies research in public administration. Finally, the results suggest that different public administration areas have different implications for disruptive technologies. The findings may be of benefit to not only the scientific community to serve as an important source for detecting associated research gaps but also to evidence-based policymaking to fully address the issues related to disruptive technologies in public administration in the future.

Abstract: In this paper, several LSTM model structures were developed, analyzed, and evaluated for being deployed to predict the future state of the CNC machine tools, one step ahead in terms of energy consumption. For the modeling purpose, self-generated datasets regarding the energy consumption of the CNC milling machine tool were applied for model training and validation in a ratio of 80/20 respectively, while the model evaluation was performed upon the CNC lathing machine tool energy consumption dataset. Furthermore, model rankings were performed based on PROMETHEE (Preference Ranking Organization METHod of Enrichment Evaluation) method which enabled the selection of the most acceptable one in the finite set of alternatives (model structures), linked to multiple criteria considered. All of the generated models generalize the observed problem quite well and can be applied according to the task that has been set.

Abstract: Finding a job in the IT industry can be a laborious process – the saturation of the job market, along with the evergrowing number of career paths in the IT sector, is making it more difficult to filter relevant career opportunities. Compounding this, many educational sources and unclear job requirements on online recruitment platforms, introduce additional challenges in navigating the job space. In this paper, we propose a rule-based recommendation system designed to match users with the most relevant job offers and help them prepare for the interview process by suggesting educational materials. The system performs the role of a domain expert in the field of professional orientation. It suggests job position/job offer recommendations that best suit a user’s skills and career personality. The knowledge base is used by the Drools rule engine. A survey was conducted to rank the most desirable company benefits. Survey results were integrated into domain rules for ranking job offers in the user’s recommendation list, alongside the rules on how much the user fulfills the job requirements. For the part of the preparation for the job interview, the system can offer educational materials that match the appropriate IT skill level of the user.

Abstract: It is a common thing that some consequences will be introduced by the doctor's actions/inactions before, during, or after the surgery. Considering wrong dosing in anesthesia the most common mistakes are caused by doctors' fatigue and inexperience, or by insufficient information included in the dose consideration. This paper proposes software solutions for improving the current state of the models in anesthesia by allowing cooperation between doctor and expert system. Integrating expert knowledge and experience within the rule-based system to guide doctor decisions can produce a safer dosing result, and reduce the percentage of occurrence of side effects. The system recommends the dosage in anesthesia and gives all relevant parameters for the displayed set of the models. The system is based on rules, which can be changed and supplemented, while the system is in production. In the current state, the proposed application can be thought of as a support system for the anesthesiologist, not their replacement.

Abstract: Open data is a concept in which specific data should be freely available to all, for use and re-use, without copyright or other restrictions. The term open data most often refers to tabular and textual data, which state institutions create. This paper will describe an open data web portal developed at the School of Electrical Engineering - University of Belgrade (ETF), with the idea to help organize audio files of the Laboratory for Acoustics. The platform has developed in a modern microservice architecture, which is used in complex systems that contain a huge number of functionalities and big data. The initial goal of the platform was to facilitate the daily activities of employees within the laboratory, but it was later concluded that the addition of new microservices could develop a portal that could be used publicly for a very wide range of people, from researchers and engineers to musicians and students. This platform allows users to upload and download available audio files, to comment and rate other users' files, as well as exchange messages between users.

Abstract: The global fleet of electric and hybrid vehicles (EVs) is predicted to grow immensely over the next decade, leading to lower CO2 emissions in road transportation but higher demand for lithium-ion batteries. Battery manufacturing and even battery recycling are both very carbon and energy-intensive processes. Having environmental sustainability in mind, reusing is the preferable technique to the production of waste and recycling since it provides an opportunity to extend the EVs batteries’ lifespan by reusing them in different second-life applications. Reusing batteries in battery energy storage systems (BESS) complements the idea of a smart grid by allowing energy storage at periods of low demand at night and release during the grid peaks, grid frequency regulation and levelling peaks in renewable energy generation, thereby alleviating the intermittent nature of renewable energy sources and decreasing the carbon footprint of the energy sector. The aim of this paper is to conduct an analysis on the benefits of reusing and repurposing EV batteries in smart BESS compared to direct recycling by using available data and case studies. A careful review of the materials shows that the benefits outweigh the disadvantages of repurposing vehicle batteries into secondary applications considering the carbon-intensiveness of the battery recycling process. The benefits of using repurposed battery packs have a major potential to facilitate a greener energy sector by shifting electricity purchases to off-peak times, more efficient use of the energy grid by providing constant energy reserves and storage for meeting changes in supply and demand, emission reductions, integration of renewable power and supporting their expansion, as well as more effective use of original materials.

Abstract: Innovations in blockchain technologies are expected to have a significant effect on many industry fields. In this paper, we explore the position of Domain Specific Languages (DSLs) in today's world, especially in emerging P2P energy trading powered by blockchain. Our motivation to research these areas is the potential of P2P trading in blockchain to revolutionize the energy sector. Existing programming languages cannot solve all the issues when working with smart contracts in energy blockchain. We elaborate usage of JetBrains MPS and KernelF as a core for developing a mini-DSL for P2P energy trading. We put forward the idea that DSLs based on KernelF can overcome shortcomings in employing existing programming languages in the given domain.