Background: Deep Acute pancreatitis (AP) is an urging cause of hospitalization in the gastroenterology due to different causes and an unpredictable outcome. Known causes are grouped into four main groups: metabolic, mechanical, vascular and infectious. Objective: To determine the role of certain biochemical or radiological parameters as predictors of an involvement of other organs in AP different pathological staging and the surgical outcome in the treatment of AP. Methods: Ninety-seven AP patients hospitalized in General Hospital “Prim.dr Abdulah Nakaš” Sarajevo, in a period between 2016 and 2021 for both sexes, were divided according to the etiological factors of AP into four groups: nutritional factors, biliary concernments, alcohol and morphological changes of the pancreas. Beside laboratory tests, the imaging methods of abdomen (transabdominal ultrasound, abdominal computed tomography) used in determining morphological changes in the pancreas and other organs were analyzed in relation to parameters that predict the need for surgical outcomes. Results: AP etiological factors of patients differ significantly by gender and showed the dominance of dietary factors in female subjects (51%), followed by the presence of concernments in the biliary tract in 36% of cases, and alcohol consumption in male subjects in 28% of cases. The only variable correlated with the indicator of necessity for surgery is the existence of pleural effusion (coefficient of correlation was 0.38; risk ratio was 5.5) resulting that patients with pleural effusion have a 5.5 times higher chance of surgery indication than other patients. Conclusion: The application of simple parameters such as creatinine value with the values of amylases in serum and urine and the presence of pleural effusion confirmed by radiological imaging of the lungs opens the possibility of a simple and effective selection of patients for surgical treatment with a more severe form of AP.

Today's extensive requirements for the storage, management, and analysis of complex, dynamic, evolving, distributed, and heterogeneous data from different sources and platforms, e.g., Big data, generate enormous challenges for IT, especially database applications. That is why the demand for data reduction is increasingly coming from the world of databases, intending to reduce the costs of storing, processing, and querying Big data. There is a large number of different techniques for Big data reduction that can cause confusion and complicate this process. Because of that, the authors proposed a Big data reduction framework to structure and present both data reduction techniques and necessary components essential for a better understanding of the process. The importance and the components of the proposed framework are explained in this paper.

Background: Non-alcoholic fatty liver disease (NAFLD) is an increasingly common cause of chronic liver disease and is becoming a major public health problem. NAFLD has been recognized as a hepatic manifestation of metabolic syndrome, associated with systemic diseases such as cardiovascular disease (CVD) and chronic kidney disease (CKD). Objective: The aim of this study was to examine the role of serum LFT parameters and renal function parameters as predictors of unmanifested liver disease. Methods: In this study, the presence of possible liver disease detected by biochemical parameters and confirmed by Transient Liver Elastography (TE) in a group of patients with different stages of chronic kidney disease (CKD) was investigated. Patients with various stages of CKD were divided into five subgroups regarding aetiology: nephroangiosclerosis, diabetic nephropathy, glomerulonephritis and pyelonephritis, autoimmune kidney disease, and polycystic and another morphological kidney disease. Liver stiffness was used to quantify liver fibrosis while Controlled attenuation parameter (CAP) was used to quantify liver steatosis. Functional liver tests and biochemical parameters of kidney function were measured in all patients. Results: Statistical analysis used in this study was a decision tree as a predictive model to map observed variables resulting in the conclusion about outcomes. The application of existing laboratory parameters, in combination with other parameters in presence of the defined etiological factors of kidneys diseases, indicate development of hepatic diseases. Higher values of phosphorus and low values of ferritin in patients with autoimmune kidney disease, and polycystic and another morphological kidney disease, expresses steatosis of the hepatic parenchyma. Conclusion: In contrary, low values of phosphorus and higher values of ferritin in patients with nephroangiosclerosis, diabetic nephropathy, glomerulonephritis and pyelonephritis, are in a favour steatosis of the hepatic parenchyma. Serum values of phosphorus and ferritin are valuable predictors of the liver disease in patients with end-stage kidney diseases of different aetiology.

Healthcare information systems store a huge amount of patient data, so the trend of the use of data mining in healthcare is on the rise. Heart and blood vessel diseases are a leading cause of mortality both worldwide and here in Bosnia and Herzegovina, and prevention, surveillance and treatment are of great public health importance. Based on data on patients with cardiovascular disease, collected from 2011 to 2017 at Mostar Hospital, models for mortality prediction using techniques for data tree mining, neural network and logistic regression are presented. The aim of this research is to compare the effectiveness of these methods in modeling the effectiveness of predicting mortality in patients with cardiovascular disease.

The article presents research about the use of genetic algorithms in the analysis of the interrelation among curriculum courses in higher education. The authors used genetic algorithms as a method to analyse the influence that achieved grades in predictors' courses have on achieved grades in dependent courses as well as to observe whether the genetic algorithms can contribute to improving the curriculum. The research was based on a set of data related to the success of students from the Faculty of Information Technologies at the University 'Džemal Bijediae' in Mostar, Bosnia and Herzegovina. The aim was to anticipate students' grades based on the grades they obtained in previous semester's courses. This research should help educational institutions to evaluate the suitability of the sequence of courses within the curriculum in order to enable personalized learning paths, make the teaching processes more efficient, and promote a balanced curriculum. Namely, a good curriculum can attract new students, improve the success rate of enrolled students, and increase the quality and visibility of the institution. Since the genetic algorithm is search techniques for handling complex spaces, we can use it for the research at each stage of the educational process. Analyses of quantitative data using a genetic algorithm can help educational institutions improve the quality of teaching.

Clostridioides difficile (CD) is a spore-forming bacterium that causes life-threatening intestinal infections in humans. Although formerly regarded as exclusively nosocomial, there is increasing genomic evidence that person-to-person transmission accounts for only <25% of cases, supporting the culture-based hypothesis that foods may be routine sources of CD-spore ingestion in humans. To synthesize the evidence on the risk of CD exposure via foods, we conducted a systematic review and meta-analysis of studies reporting the culture prevalence of CD in foods between January 1981 and November 2019. Meta-analyses, risk-ratio estimates, and meta-regression were used to estimate weighed-prevalence across studies and food types to identify laboratory and geographical sources of heterogeneity. In total, 21886 food samples were tested for CD between 1981 and 2019 (96.4%, n = 21084, 2007–2019; 232 food-sample-sets; 79 studies; 25 countries). Culture methodology, sample size and type, region, and latitude were sources of heterogeneity (p < 0.05). Although non-strictly-anaerobic methods were reported in some studies, and we confirmed experimentally that improper anaerobiosis of media/sample-handling affects CD recovery in agar (Fisher, p < 0.01), most studies (>72%) employed the same (one-of-six) culture strategy. Because the prevalence was also meta-analytically similar across six culture strategies reported, all studies were integrated using three meta-analytical methods. At the study level (n = 79), the four-decade global cumulative-prevalence of CD in the human diet was 4.1% (95%CI = −3.71, 11.91). At the food-set level (n = 232, mean 12.9 g/sample, similar across regions p > 0.2; 95%CI = 9.7–16.2), the weighted prevalence ranged between 4.5% (95%CI = 3–6%; all studies) and 8% (95%CI = 7–8%; only CD-positive-studies). Risk-ratio ranking and meta-regression showed that milk was the least likely source of CD, while seafood, leafy green vegetables, pork, and poultry carried higher risks (p < 0.05). Across regions, the risk of CD in foods for foodborne exposure reproducibly decreased with Earth latitude (p < 0.001). In conclusion, CD in the human diet is a global non-random-source of foodborne exposure that occurs independently of laboratory culture methods, across regions, and at a variable level depending on food type and latitude. The latitudinal trend (high CD-food-prevalence toward tropic) is unexpectedly inverse to the epidemiological observations of CD-infections in humans (frequent in temperate regions). Findings suggest the plausible hypothesis that ecologically-richer microbiomes in the tropic might protect against intestinal CD colonization/infections despite CD ingestion.

The aim of this research is to create fully functional environment for real-life testing of various algorithms in mobile robotic. Main advantage, beside low-cost, is dealing with challenges of real-life implementation. Unlike simulation environments (e.g. MathLab), this setting will allow researchers to test their path planning, collision detection and other algorithms with real challenges, real robots and real static and dynamic obstacles. Environment is particularly adapted for swarm robotics and any other mobile robotics including 2D and 3D scenarios. Enabling researches to prove their algorithms in this environment allows significantly faster path to real implementation in industry, military and science. Concept is based on OpenCV library and low-cost hardware. Solution is based on visual positioning and motion vector detection. System allows controlling mobile robots via radio communication with the range up to 100m which allows practical application of the system. This research paper should be considered as a part of series of research papers published earlier.

What people say on social media has turned into a rich source of information to understand social behavior. Specifically, the growing use of Twitter social media for political communication has arisen high opportunities to know the opinion of large numbers of politically active individuals in real time and predict the global political tendencies of a specific country. It has led to an increasing body of research on this topic. The majority of these studies have been focused on polarized political contexts characterized by only two alternatives. Unlike them, this paper tackles the challenge of forecasting Spanish political trends, characterized by multiple political parties, by means of analyzing the Twitters Users political tendency. According to this, a new strategy, named Tweets Analysis Strategy (TAS), is proposed. This is based on analyzing the users tweets by means of discovering its sentiment (positive, negative or neutral) and classifying them according to the political party they support. From this individual political tendency, the global political prediction for each political party is calculated. In order to do this, two different strategies for analyzing the sentiment analysis are proposed: one is based on Positive and Negative words Matching (PNM) and the second one is based on a Neural Networks Strategy (NNS). The complete TAS strategy has been performed in a Big-Data environment. The experimental results presented in this paper reveal that NNS strategy performs much better than PNM strategy to analyze the tweet sentiment. In addition, this research analyzes the viability of the TAS strategy to obtain the global trend in a political context make up by multiple parties with an error lower than 23%. Keywords—Political tendency, prediction, sentiment analysis, Twitter.