Search

Inability of safety managers to timely detect misuse of Personal protective equipment (PPE) causes a number of injuries and financial losses. Considering sizes of industry halls and number of workers, there is an increasing demand for computerized tools that could help companies to enhance the implementation of strictinging workplace safety standards. As a solution, we propose a procedure that: 1) reduces the problem of PPE compliance to the binary classification, and 2) enables compliance of arbitrary type and number of PPE that could be mounted on various body parts. To prove this hypothesis, we studied 18 different PPE types used across various industries for protecting 5 physiological body parts/functions. The HigherHRNet pose estimator was used for defining the PPE regions of interest, while six different image classification architectures were assessed for the compliance/classification of the considered regions. All classifiers were pretrained on the ImageNet data set and fine-tuned using the dedicated data set developed during this study. Top-performing models were MobileNetV2, Dense-Net, and ResNet, while the MobileNetV2 was recommended as the most optimal choice considering its lower computation demands. Compared to previous studies, the proposed approach demonstrated competing performances with unique ability to be easily adopted for performing compliance of various PPE by slight editing of the predefined lists of PPE types and corresponding body parts. Considering the present data/privacy/ computational constraints, the procedure is recommended as suited for the digitalization of PPE compliance in: 1) self-check points, and 2) safety-critical workplaces.

Nowadays, the sovereign credit rating is not only an index of a country's economic performance and political stability but also an overall indicator of development and growth, as well as the trust factor that is associated with the country. Due to its importance, the vast amount of available information, and the lack of a closed-form solution, prediction models based on machine learning (ML) and computation intelligence (CI) techniques are being increasingly used to complement traditional financial approaches. In this paper, we aim to introduce a novel ML-CI approach for sovereign credit rating prediction based on a differential evolution (DE) algorithm and interpolative Boolean algebra (IBA). In fact, the proposed approach is based on a pseudo-logical function in the IBA framework derived from the historical data of publicly available indicators using the DE algorithm. Such functions are easily interpreted and enable a subtle gradation among countries. It is shown that the IBA-DE approach outperforms back-propagation neural networks on the observed problem while also providing a deeper insight into each of the indicators used for prediction and its respective influence on the prediction rating on the other.

One of the main issues in IFS theory are generalizations of intuitionistic fuzzy set (IFS) definition as well as IFS operations. In this paper, we present the LBIFS-IBA approach by applying operations based on interpolative Boolean algebra (IBA) on generalized IFS. Namely, LBIFS are defined as a special case of Liu's generalized IFS with the maximal interpretational surface. By extending the interpretational surface, the descriptive power of the approach is enhanced, and therefore the problematic situations when (mu)A+nu(A)>1 can be modeled. In addition, IBA-based algebra secures Boolean properties of the proposed approach. Considerable attention is given to comprehension of uncertainty within LBIFS-IBA, i.e., we propose a novel manner of uncertainty interpretation by treating values from [-1,1] interval. In order to prove its importance, we compare LBIFS-IBA with several well-known IFS generalizations, showing that only our approach offers meaningful uncertainty interpretation is all selected cases. Additionally, we illustrate the practical benefits of LBIFS-IBA by applying it to an example of modeling Japanese candlesticks for price charting and paying special attention to uncertainty interpretation.

Airline decision about how many seats to allow to be overbooked is based on the expectation of the number of passengers who will not show up on a specific flight. This paper proposes a decision support system for predicting the number of no show passengers that combines the case-based reasoning (CBR) approach with Interpolative Boolean Algebra (IBA) and considers recommendations from both expert and algorithm. More precisely, recently proposed IBA similarity measure along with suitable aggregation operator is used for comparing alternatives in CBR algorithms. The proposed system was tested on the real-life data of the Belgrade-Amsterdam route. The obtained results show the necessity to include expert knowledge in the prediction process. Furthermore, the results are indicating that IBA-based models perform significantly better comparing to traditional distance-based models. The proposed expert system should contribute to an airline utilizing its inventory, which will further result in profit increase.

Workplace safety is a scientific discipline that has been constantly evolving along with industrial development. Nowadays technological progress of tools and materials used in industry, in addition to all the positive impacts, increase the probability of injuries of the operators that use them. Consequently, there are industry standards and recommendations that specify appropriate personal protective equipment (PPE) for certain workplaces. Although every company is able to provide protective equipment for its employees, the major challenge is the compliance and control of their proper use. The aim of this study was to assess the possibility of applying artificial intelligence and deep learning techniques for automated PPE compliance, which could help in taking preventive action with the aim of reducing injuries caused due to non-use or misuse of prescribed PPEs. The obtained results showed that the YOLOv5 algorithm achieved high precision (average 0.857) for the detection of various types of head-mounted personal protective equipment. Accordingly, there is a high potential for future use of such tools in improving workplace safety and PPE compliance. Potential users of the application based on this recognition algorithm would be companies which regulations define the type of PPEs that have to be used at a certain working position.

Recommender systems help users to reduce the amount of time they spend to find the items they are interested in. One of the most successful approaches is collaborative filtering. The main feature of a recommender system is its ability to predict user’s interests by analyzing the behavior of this particular user and/or the behavior of other similar users to generate personalized recommendations. Identification of neighbor users who have had similar taste to the target user in the past is a crucial process for successful application of collaborative filtering. In this paper, we proposed a collaborative filtering method that uses interpolative Boolean algebra for calculation of similarity between users. In order to analyze the effectiveness of the proposed approach we used three common datasets: MovieLens 100K, MovieLens 1M, and CiaoDVD. We compared a collaborative filtering based on IBA similarity measure with two standard similarity measures: Pearson correlation and cosine-based coefficient. Even though statistical measures are traditionally used in recommender systems, proposed logic-based approach showed promising results on the tested datasets. A recommender system with IBA similarity measure outperformed the others in most cases.

The idea of this paper is to investigate intuitionistic fuzzy set (IFS) generalizations and make one step further into that direction by introducing interpolative Boolean algebra (IBA) to Liu's generalized IFS. First, we aim to give an overview of the existing attempts to generalize or alter the definition of IFS and to underline the most general approach. Further, Liu's bipolar I-fuzzy sets (LBIFSs) are introduced as the most universal case of Liu's generalized intuitionistic fuzzy sets. Finally, IBA-based operations are introduced to deal with LBIFS. In the proposed LBIFS-IBA approach, all Boolean laws are fulfilled. The special attention is dedicated to laws of the excluded middle and contradiction.

In this paper we propose a new approach for forecasting the cryptocurrency time series, which combines the fuzzy transform and the fuzzy inference system. We also test whether fuzzy transform yields better results forecasting results in comparison to Fourier transform. Finally, we will investigate whether fuzzy rules used in fuzzy inference system can successfully capture high and low volatility moments in the time series, as well as high correlations between the three different cryptocurrencies.

Problem prepoznavanja lica se u poslednjih dvadeset godina smatra jednim od najaktuelnijih i najbitnijih. Ovaj problem suštinski se svodi na probleme identifikacije i verifikacije osobe sa digitalne slike ili video signala. U ovom radu veštačke neuronske mreže i metoda nosećih vektora primenjene su kao klasifikatori pri rešavanju problema identifikacije lica. Dalje, ispitivano je koji algoritam ekstrakcije podataka iz slike je bolje koristiti uz koji klasifikator. Na kraju, analizirano je da li normalizacija osvetljenja ima značajan uticaj na rezultate identifikacije. Eksperiment je izvršen na često korišćenoj AT&T bazi slika osoba. Iako su dobijeni zadovoljavajući rezultati za obe klasifikacione metode, metoda nosećih vektora se pokazala kao nešto bolja na ovom problemu., The face recognition problem is one of the most challenging problems in the last twenty years. This problem may come down on identifying or verifying a person from a digital image or a video frame. In this paper, artificial neural networks and support vector machines are used as classifiers to solve a problem of face identification. Further, we aim to investigate which feature extraction algorithm is more suitable to be used along with which classifier. Finally, we analyze if the illumination normalization has a significant influence to identification results. The experiment is performed on the well-known AT&T image database. Although the obtained results are satisfactory for both classification methods, support vector machines proved to be slightly better for solving this problem.

U ovom radu se analiziraju deskriptivne karakteristike arbitražnih prilika koje se javljaju nad valutnim parovima, i ispituju mogućnosti praktične realizacije triangularnih arbitražnih strategija na međunarodnom deviznom tržištu. Cilj rada je da se odgovori na pitanje da li je moguće, i pod kojim uslovima, iskoristiti triangularnu arbitražu na tržištu. U tu svrhu razvijen je automatizovani sistem za trgovanje. Predloženi sistem je sastavljen iz više komponenti pri čemu je svaka karakteristika arbitraže opisana pomoću jedne komponente sistema. Nakon prikaza predloženog sistema, u radu su objašnjene i testirane različite strategije za trgovanje. Rezultati simulacija prikazani su zajedno sa analizom isplativosti navedenih strategija i uslovima pod kojima bi one mogle da ostvare profit. Za potrebe testiranja korišćeni su podaci o kretanju deviznog kursa za odabrane valutne parove u toku 2012. i 2013. godine., The aim of this paper is to examine whether it is possible to seize the triangular arbitrage opportunities in the foreign exchange market, and under what conditions. First the descriptive characteristics of arbitrage opportunities that occur between the currency pairs are analyzed, and further the profitability of the triangular arbitrage strategies is investigated. For that purpose, an automated trading system is developed. The proposed system is composed of several components where each one provides analysis for a specific arbitrage characteristic. After the system structure is given, a set of trading strategies is described and further tested. The simulation results are presented along with the analysis of profitability for the tested strategies and conditions under which they might make a profit. The arbitrage strategies are simulated for the selected currency pairs based on their exchange rate movement in the year 2012 and 2013.

In this paper, we present logic-based aggregation models used for ranking student applicants and we compare them with a number of existing aggregation methods, each more complex than the previous one. The proposed models aim to include dependencies in the data using Logical aggregation (LA). LA is a aggregation method based on interpolative Boolean algebra (IBA), a consistent multi-valued realization of Boolean algebra. This technique is used for a Boolean consistent aggregation of attributes that are logically dependent. The comparison is performed in the case of student applicants for master programs at the University of Belgrade. We have shown that LA has some advantages over other presented aggregation methods. The software realization of all applied aggregation methods is also provided. This paper may be of interest not only for student ranking, but also for similar problems of ranking people e.g. employees, team members, etc.

Boolean networks are used for modeling and analysis of complex systems of interacting entities. Classical Boolean networks are binary and they are relevant for modeling systems with complex switch-like causal interactions. More descriptive power can be provided by the introduction of gradation in this model. If this is accomplished by using conventional fuzzy logics, the generalized model cannot secure the Boolean frame. Consequently, the validity of themodel's dynamics is not secured. The aim of this paper is to present the Boolean consistent generalization of Boolean networks, interpolative Boolean networks. The generalization is based on interpolative Boolean algebra, the [0, 1]-valued realization of Boolean algebra. The proposed model is adaptive with respect to the nature of input variables and it offers greater descriptive power as compared with traditional models. For illustrative purposes, IBN is compared to themodels based on existing real-valued approaches. Due to the complexity of themost systems to be analyzed and the characteristics of interpolative Boolean algebra, the software support is developed to provide graphical and numerical tools for complex system modeling and analysis.

In this paper, we propose a decision support system for enhancing scrum methodology based on fuzzy logic. Scrum is a very popular agile methodology within the software and product development. In the basic scrum, requirements that describe a certain task do not have a clear interpretation. Also, the traditional model does not take into account the experience of the developers nor the logical dependencies of input variables. Fuzzy inference is particularly useful for this purpose, because it incorporates logic in inference process and inputs are presented using linguistic quantifiers. The proposed system consists of three main components: a fuzzy inference system, an aggregation operator and a feedback function. The aggregation function is used to aggregate task predictions in a single value that uniquely represent a specific task, while a feedback is employed to adjust an input variable to improve system performance. Furthermore, the proposed system is simulated with randomly generated inputs in order to analyse its behaviour. The predictions of the system are more accurate and with smaller deviation in the final iterations.

Акутни панкреатитис је ензиматско инфламаторно оболење панкреаса, са инциденцијом око 17/100000 становника. Најчешћи етиолошки чиниоци који се везују за ово стање су билијарна калкулоза (45%) и конзумација алкохолних пића (35%). Ређи узроци су одређени лекови, хипертриглицеридемија, хиперкалијемија, траума, урођени чиниоци, и идиопатски панкреатитис (20%). По међународној Атланској (Atlanta) класификацији, акутни панкреатитис се може манифестовати у умереној, умерено тешкој и тешкој форми. Умерене форме панкреатитиса се јављају у 80%, карактеришу се едемом органа и имају благ и краткотрајан клинички ток, са стопом морталитета од 1%. Основни циљ лечења болесника са акутним панкреатитисом у прва 24 сата хоспитализације јесте олакшавање тегоба, утврђивање узрока панкреатитиса и процена тежине обољења. Акутни панкреатитис билијарне етиологије узрокован је калкулозом жучне кесе и/ или жучних путева. Препоручени третман билијарне калкулозе, у циљу превенције поновног атака умерене форме билијарног панкреатитиса , подразумева уклањање жучне кесе лапароскопском холецистектомијом са интраоперативном холангиографијом. Уколико се дијагностикује калкулоза жучних канала ради се ендоскопска ретроградна холангиопанкреатографија (ЕРЦП) са ендоскопском папилотомијом (ЕПТ) и уклањем__ калкулуса и детритуса уз жучних водова, са циљем обезбеђивања нормалног протока жучи у дванаестопалачно црево. Циљ овог истраживања је оптимализација редоследа примене ЛХ и ЕРЦП са ЕПТ, идентификацијом предикционих показатеља холедохолитијазе, ради скраћења дужине хоспитализације болесника са умереном формом акутног билијарног панкреатитиса. У спроведеној проспективној анамнестичкој студији, учествовало је 100 болесника лечених од умерене форме акутног билијарног панкреатитиса, у Ургентном центру Клиничког центра Војводине, од 2011. до 2015.године, од којих је код 80 урађена само ЛХ, а код 20 ЛХ и ЕРЦП са ЕПТ. Анализом клиничких, ултразвучних и лабораторијских налаза, идентификовано је 5 статистички зна, Akutni pankreatitis je enzimatsko inflamatorno obolenje pankreasa, sa incidencijom oko 17/100000 stanovnika. Najčešći etiološki činioci koji se vezuju za ovo stanje su bilijarna kalkuloza (45%) i konzumacija alkoholnih pića (35%). Ređi uzroci su određeni lekovi, hipertrigliceridemija, hiperkalijemija, trauma, urođeni činioci, i idiopatski pankreatitis (20%). Po međunarodnoj Atlanskoj (Atlanta) klasifikaciji, akutni pankreatitis se može manifestovati u umerenoj, umereno teškoj i teškoj formi. Umerene forme pankreatitisa se javljaju u 80%, karakterišu se edemom organa i imaju blag i kratkotrajan klinički tok, sa stopom mortaliteta od 1%. Osnovni cilj lečenja bolesnika sa akutnim pankreatitisom u prva 24 sata hospitalizacije jeste olakšavanje tegoba, utvrđivanje uzroka pankreatitisa i procena težine oboljenja. Akutni pankreatitis bilijarne etiologije uzrokovan je kalkulozom žučne kese i/ ili žučnih puteva. Preporučeni tretman bilijarne kalkuloze, u cilju prevencije ponovnog ataka umerene forme bilijarnog pankreatitisa , podrazumeva uklanjanje žučne kese laparoskopskom holecistektomijom sa intraoperativnom holangiografijom. Ukoliko se dijagnostikuje kalkuloza žučnih kanala radi se endoskopska retrogradna holangiopankreatografija (ERCP) sa endoskopskom papilotomijom (EPT) i uklanjem__ kalkulusa i detritusa uz žučnih vodova, sa ciljem obezbeđivanja normalnog protoka žuči u dvanaestopalačno crevo. Cilj ovog istraživanja je optimalizacija redosleda primene LH i ERCP sa EPT, identifikacijom predikcionih pokazatelja holedoholitijaze, radi skraćenja dužine hospitalizacije bolesnika sa umerenom formom akutnog bilijarnog pankreatitisa. U sprovedenoj prospektivnoj anamnestičkoj studiji, učestvovalo je 100 bolesnika lečenih od umerene forme akutnog bilijarnog pankreatitisa, u Urgentnom centru Kliničkog centra Vojvodine, od 2011. do 2015.godine, od kojih je kod 80 urađena samo LH, a kod 20 LH i ERCP sa EPT. Analizom kliničkih, ultrazvučnih i laboratorijskih nalaza, identifikovano j, Acute pancreatitis is an enzymatic inflammatory disease of the pancreas, with an incidence of around 17/100000 inhabitants. The most common etiological factors that are associated with this condition are biliary calculi (45%) and consumption of alcoholic beverages (35%). Less common causes include certain medications, hypertriglyceridemia, hyperkalemia, trauma, congenital factors and idiopathic pancreatitis (20%). According to the Atlanta International classification, acute pancreatitis can be manifested in a moderate, moderately severe and severe forms. Moderate forms of pancreatitis occur in 80%, characterized by pancreatic edema and have mild and short clinical course, with a mortality rate of 1%. The main goal of treatment of patients with acute pancreatitis in the first 24 hours of hospitalization is to facilitate complaints, determining the cause of pancreatitis and assessment of severity of the disease. Acute biliary pancreatitis is caused by calculosis of the gallbladder and / or bile ducts. The recommended treatment of biliary calculi, in order to prevent repeated attacks of moderate biliary pancreatitis, involves the removal of the gallbladder thru laparoscopic cholecystectomy with intraoperative cholangiography. If presence of bile duct calculi is established, an endoscopic retrograde cholangiopancreatography (ERCP) with endoscopic papillotomy (EPT) and removes stones and detritus along the bile ducts is indicated, with the aim of ensuring the normal flow of bile into the duodenum. The aim of this study is the optimization of the order of application LH and ERCP with EPT, the identification of predictable indicators of choledocholithiasis, in order to shorten the length of hospitalization of patients with a moderate form of acute biliary pancreatitis. We conducted prospective case control study, with 100 patients involved, treated for moderate forms of acute biliary pancreatitis in the Emergency Center of the Clinical Center of Vojvodina, from 2011 to 201