RFOS - Repository of Faculty of Organizational Sciences Univ. of Belgrade
Not a member yet
2871 research outputs found
Sort by
Collaborative Economy Index: a quest for robust and impartial composite indicator of collaborative economy
No full textCollaborative, platform, sharing. All of these terms try po encapsulate similar phenomenon: collaborative consumption that pertains to non-traditional consumption patterns where individuals utilize goods and services sourced from other individuals on the Internet, rather than from the market. Similar are the challenges that researchers and policy makers are faced with when trying to capsulate collaborative economy performance of certain entity (such as region or country). The process of integrating various indicators into one single value is often based on the composite indicators approach. A range of issues are raised throughout the process of composite index creation, with the choice of indicators and weighting scheme being the ones most frequently cited in the literature. Our paper will add to the growing body of papers on the topic of collaborative economy indexes and address the key slippery slopes of the process
Risk Management in the Automation of Business Processes through the Application of Robotic Process Automation
No full textIn a modern business environment with very big data flows, employee focus is crucial for success. This problem developed a need for new
technological advancements such as Robotic Process Automation (RPA). As
the RPA technology is relatively young, there is still an insufficient amount
of research about which project management techniques are the best for it.
This paper aims to recognize the key principles of risk management which
can be applied to the automation of business processes through the use of
RPA technology and to emphasize the significance of risk management for
conducting successful RPA projects
Comparative Analysis of graph-based NoSQL databases
No full textIn the introductory part of the paper, a brief historical overview of NoSQL databases, their
origin and use, reasons for introduction, advantages and disadvantages, similarities and differences
with relational databases will be described. In the next part, the focus of the paper will shift to graph-based NoSQL databases along with the main features and properties of three main categories of
NoSQL databases – Neo4j, ArangoDB and OrientDB. Special emphasis will be placed on structure,
implementation, examples of graph-based databases and use cases. It is important to single out main
representatives of graph-based NoSQL databases, with a more detailed description of functionalities
and architecture. All of the previously mentioned aims to present a comparative analysis of the
previously described databases according to the given criteria in order to summarize the results of the
analysis in the final part of the paper with recommendations and best practices
Poboljšanje objašnjivosti VI modela na primeru ocenjivanja timskog rada studenata programiranja
No full textU našem prethodnom istraživanju, ocenjivan je timski rad studenata programiranja preko metrika dobijenih iz repozitorijuma koda. Pedagoška osnova je bila rubrika za ocenjivanje timskog rada. Upotrebljene su tri metode mašinskog učenja (stabla odlučivanja, random forest i neuronske mreže) kao i fazi zaključivanje, a evaluacija je obuhvatala kvalitet klasifikacije, praktičnost i objašnjivost. Objašnjivost je bila naročito bitna jer su nastavnici želeli da znaju zašto su studenti dobili određene ocene tj. kako je VI metoda došla do njih. Jedan od zaključaka je bio da, iz aspekta objašnjivosti, upravo fazi zaključivanje (korišćenjem traga izvršenih pravila), na osnovu mišljenja nastavnika, pruža najshvatljivija objašnjenja, a zatim stabla odlučivanja (grafički prikaz stabla). Za neuronske mreže i random forest, u to vreme nisu bile raspoložive metode za objašnjavanje. Intenzivan razvoj oblasti objašnjive VI u toku poslednjih nekoliko godina doneo je različite metode koje omogućavaju ljudima (korisnicima) da shvate kako je neki VI model izveo određeni rezultat i da, na taj način, doprinesu poverenju u rezultate modela i sam model. U ovom radu, fokus je na diskusiji o mogućnostima modela objašnjive VI na navedenom primeru ocenjivanja timskog rada, sa osvrtom na praktičnost pojedinačnih metoda i prilagođenost objašnjenja krajnjim korisnicima. Konkretno, u radu se razmatraju metode kontra-činjeničnih objašnjenja, SHAP i druge vizuelizacije i porede sa „klasičnim“ metodama za objašnjavanje poteklim iz oblasti ekspertnih sistema.Knjiga apstrakat
Skladište podataka za analizu stope nezaposlenosti u Republici Srbiji
No full textIn the paper, we present a data warehouse system to analyze the
unemployment rate in the Republic of Serbia. The goal of our research is to improve the
analytical capabilities of the unemployment rate in Serbia by creating a new business
intelligence tool and predictive machine learning models. First, we discuss research
motives and the unemployment problem, and then we present the development process
of the proposed data warehouse system. The Data Warehouse Quality methodology has
been deployed to assess the quality of the data. Machine learning algorithms have been
utilized to build predictive models and gain insights into the differences in
unemployment rates between young and experienced workers. Finally, we have created
several reports to visually present the results of the proposed data analyses.U radu predstavljamo sistem skladišta podataka za analizu stope
nezaposlenosti u Republici Srbiji. Cilj našeg istraživanja je da unapredimo analitičke
mogućnosti stope nezaposlenosti u Srbiji kreiranjem novog alata poslovne inteligencije i
prediktivnih modela mašinskog učenja. Prvo razmatramo motive istraživanja i problem
nezaposlenosti, a zatim predstavljamo proces razvoja predloženog sistema skladišta
podataka. Metodologija kvaliteta skladišta podataka je primenjena za procenu kvaliteta
podataka. Algoritmi mašinskog učenja su korišćeni za izgradnju prediktivnih modela i
sticanje uvida u razlike u stopama nezaposlenosti između mladih i iskusnih radnika.
Konačno, napravili smo nekoliko izveštaja kako bismo vizuelno predstavili rezultate
predloženih analiza podataka
Развој међународних стандарда за сарадњу између универзитетских и пословних организација
No full textСарадња између академског и пословног света одувек је била тема која је привлачила пажњу обе ове стране, али већ и само непостојање значајнијих међународних или глобалних стандарда, који би је на одговарајући начин уредили, сведочи о чињеници да њен потенцијал још увек остаје недовољно искоришћен. У Међународној организацији за стандардизацију (ISO), у оквиру једног од активнијих техничких комитета „ISO/TC 286 Collaborative business relationship management”, релативно недавно, у марту 2022. године, оформљена је посебна радна група, са циљем да се бави развојем првог, условно речено, глобалног стандарда, којим би се дефинисале смернице за сарадњу између организација које су у склопу универзитета и свих оних осталих, привредних и непривредних, које су ван самих универзитета. Република Србија, односно њени представници – у име Института за стандардизацију Србије, пред-ставља једну од првих држава учесница у овом процесу који, организационо и администра-тивно, предводе стручњаци из Кине и Велике Британије, делегирани од стране кинеских и британских националних тела за стандардизацију. Осим упознавања са самим током пред-метног процеса, у овом раду/предавању по позиву, биће представљене и основне карактери-стике садржаја актуелне верзије, тј. тренутног нацрта стандарда „ISO 44006 Collaborative business relationship management — Guidelines for university-business collaboration”.Овај ста-ндард би, према садашњим плановима, требало да буде објављен у току 2025. године и да служи као својеврсан скуп препорука, намењен универзитетским и пословним организацијама, као и свим другим интересним и заинтересованим странама, о томе како већ постојећих 12 принципа за менаџмент добрих пословних односа, могу да примене како би побољшавали своје капацитете за међусобну сарадњу
Applying Lean Innovation in the Customs System Digitalization
No full textGlobalization, regional integration, and the era of digitalization impose the need for Customs, as an entity of the state administration, to implement process innovations in terms of integration, automation, and digitalization. Thus, the purpose of this research is to analyze the possibilities of improving the customs system by applying process innovations. For the experiment, economic operators, as end users of the customs system were employed for a case study to test and verify the proposed guidelines for the lean innovation approach. The main methods used in the research are descriptive analysis of the processed data collected by survey and the method of interview with the economic operators. The result of the research is that applying process innovation supported by lean approach prior to digitalization can significantly improve efficiency, effectiveness and shorten the execution time of customs transit procedure
UNCOMPLICATE - SOFTWARE LIBRARIES FOR MACHINE LEARNING AND HIGH PERFORMANCE COMPUTING
No full text. We present Uncomplicate, a family of functional software libraries written in Clojure
that runs on central processing units (CPU) and graphic processing units (GPU). Among others,
this includes the following libraries: Deep Diamond, a tensor-based deep neural networks library,
Neanderthal, which handles vectors, matrices, and linear algebra, and ClojureCUDA, a functional
library for GPU programming.
What sets Uncomplicate libraries apart is that they offer relevant constructs at various levels of
abstraction, from low-level bindings to the high-level domain-specific code, without
compromising neither elegance nor performance. Despite being written in Clojure, a high-level
dialect of Lisp, they offer full performance of the state of the art lowest level operations provided
by the underlying vendor libraries of low-level operations such as Nvidia’s cuDNN (CUDA Deep
Neural Networks) and Intel’s MKL (Math Kernel Library), on CPU as well as GPU.
The implementation details of all Uncomplicate libraries are freely accessible online at
https://github.com/uncomplicate, including the whole history of development in Git
repositories. All libraries are available as free software under Eclipse Public License.
Compared to mainstream alternatives, such as Python’s NumPy, TensorFlow, Pytorch, and similar
libraries, we achieved comparable and better computation performance (where applicable). The
drawback is, of course, that we cannot even approach the popularity and invested resources that
the Python ecosystem receives. On the other hand, Uncomplicate libraries cover more
functionality (for features that are applicable) with orders of magnitude less code. As an example,
Neanderthal, a rough match to NumPy, is only 20,000 lines of code in total compared to NumPy’s
400,000. In addition to that, Neanderthal supports CPU, CUDA GPU, OpenCL GPU, and OpenCL
CPU execution, while NumPy’s core implementation only supports running on CPU. Uncomplicate
libraries are implemented more effectively, backed by the fact that it has been developed by one
developer in free time after work, while mainstream equivalents are developed by multitude of
full-time employees. It is a sort of an apples to oranges comparison, we admit, but we are open
to reviews and further suggestions
Fluorescence spectra of aflatoxin in cow’s milk and coffee
No full textThis article considers the fluorescence spectra of aflatoxin in cow’s milk and coffee grounds. It was shown that fluorescence spectra of aflatoxin change in time. Fluorescence maximum I increases linearly with time and can be expressed with linear regression. For cow’s milk, regression is I[rel.unit]=195+17.75´t [days] and for coffee grounds it is I[rel.unit]= 240+9.58´t(days). This work presents a small step towards better protection of human health from the unfavorable effect of aflatoxin
CROWD-BASED OPEN INNOVATION: MODELS, CHALLENGES, AND TRENDS
No full textThis article considers concepts, models, achieved results and emerging trends of the open innovation approach to organizing and conducting research and development activities in companies. The goal is to perform a comprehensive systematic analysis of literature that considers the open innovation models that focus on the collaboration of various stakeholders, including companies, academia, government, startups, individuals, and others, in the context of design and development of innovative digital services, through integration of crowdsourcing and DevOps. As a result, we propose a new framework for organizing open innovation activities using DevOps practices for digitals service development. Finally, we present exampl