A5045389796- Machine Learning and Data Classification
- Data Stream Mining Techniques
- Anomaly Detection Techniques and Applications
- Advanced Statistical Methods and Models
- Imbalanced Data Classification Techniques
- Topic Modeling
- Cardiovascular Function and Risk Factors
- Natural Language Processing Techniques
- Recommender Systems and Techniques
- Data Mining Algorithms and Applications
- Time Series Analysis and Forecasting
- Cardiomyopathy and Myosin Studies
- Artificial Intelligence in Healthcare
- Complex Network Analysis Techniques
- Dementia and Cognitive Impairment Research
- Statistical Distribution Estimation and Applications
- Fault Detection and Control Systems
- Probabilistic and Robust Engineering Design
- Educational Technology and Assessment
- Machine Learning and Algorithms
- Machine Learning in Healthcare
- Software Engineering Research
- Domain Adaptation and Few-Shot Learning
- Advanced Malware Detection Techniques
- Security and Verification in Computing
University of Ljubljana
2015-2025
Ball State University
2023
VA Palo Alto Health Care System
2023
Stanford University
2023
University of Belgrade
2023
University of Illinois Chicago
2023
Centar za Promociju Nauke
2023
Mayo Clinic
2023
University of Osijek
2023
Ljubljana Passenger Transport
2022
This paper proposes a system for the early automatic recognition of health problems that manifest themselves in distinctive form gait. Purpose is to prolong autonomous living elderly at home. When identifies problem, it automatically notifies physician and provides an explanation diagnosis. The gait user captured using motion-capture system, which consists body-worn tags wall-mounted sensors. positions are acquired by sensors resulting time series position coordinates analyzed with...
The use of ROC (Receiver Operating Characteristics) analysis as a tool for evaluating the performance classification models in machine learning has been increasing last decade. Among most notable advances this area are extension two-class to multi-class case well employment cost-sensitive learning. Methods now exist which take instance-varying costs into account. purpose our paper is present survey field with aim gathering important achievements one place. In paper, we application areas...
Machine learning (ML) and artificial intelligence are emerging as important components of precision medicine that enhance diagnosis risk stratification. Risk stratification tools for hypertrophic cardiomyopathy (HCM) exist, but they based on traditional statistical methods. The aim was to develop a novel machine tool the prediction 5-year in HCM. goal determine if its predictive accuracy is higher than state-of-the-art tools. Data from total 2302 patients were used. data comprised...
The vastness of chemical space and the relatively small coverage by experimental data recording molecular properties require us to identify subspaces, or domains, for which we can confidently apply QSAR models. prediction models in these domains is reliable, potential subsequent investigations such compounds would find that predictions closely match values. Standard approaches assume are more reliable "similar" those subspaces with denser data. Here, report on a study an alternative set...
In Machine Learning, estimation of the predictive accuracy for a given model is most commonly approached by analyzing average model. general, models do not provide estimates their individual predictions. The reliability pred ictions require analysis various and instance properties. paper we make an overview approaches prediction reliability. We start summarizing three research fields, that provided ideas motivation our work: (a) to perturbing learning data, (b) usage unlabeled data in...
Introduction Heart failure (HF) is a complex clinical syndrome. Accurate risk stratification and early diagnosis of HF are challenging as its signs symptoms non-specific. We propose to address this global challenge by developing the STRATIFYHF artificial intelligence-driven decision support system (DSS), which uses novel analytical methods in determining risk, prognosis HF. The primary aim present study collect prospective data validate DSS (in terms diagnostic accuracy, sensitivity...
In the paper, we present an empirical evaluation of five feature selection methods: ReliefF, random forest selector, sequential forward selection, backward and Gini index. Among evaluated methods, f
Extracorporeal hemadsorption may reduce inflammatory reaction in cardiopulmonary bypass (CPB) surgery. Glucocorticoids have been used during open-heart surgery for alleviation of systemic inflammation after CPB. We compared intraoperative and methylprednisolone, with usual care, complex cardiac on CPB, responses, hemodynamics, perioperative course. Seventy-six patients prolonged CPB were recruited randomized, 60 included final analysis. Allocation was into three groups: Methylprednisolone (n...
The paper presents an approach to the task of automatic document categorization in field economics. Since documents can be annotated with multiple keywords (labels), we this by applying and evaluating multi-label classification methods supervised machine learning. We describe forming a test corpus 1015 economic that automatically classify using tool which integrates ontology construction text mining methods. In our experimental work, evaluate three groups approaches: transformation...
Objective The consumption of opioid analgesics could be reduced by the use with different mechanisms action. We investigated whether additional treatment dexmedetomidine or lidocaine reduce consumption. Methods randomized 59 study participants into three groups and examined: (i) fentanyl consumption, (ii) piritramide, (iii) cognitive function neuropathic pain. control group received continuous propofol infusion boluses. Continuous intravenous (0.5 µg/kg/h) was administered to (1.5 mg/kg/h)...
Cardiovascular disorders in general are responsible for 30% of deaths worldwide. Among them, hypertrophic cardiomyopathy (HCM) is a genetic cardiac disease that present about 1 500 young adults and can cause sudden death (SCD).Although the current state-of-the-art methods model risk SCD patients, to best our knowledge, no available modeling patient's clinical status up 10 years ahead. In this paper, we propose novel machine learning (ML)-based tool predicting progression patients diagnosed...
Abstract When programmers need to modify third‐party applications, they frequently do not have access their source code. In such cases, DLL injection and API hooking are techniques that can be used applications without intervening into The commonly varieties of approaches many practical limitations: inconvenient for a programmer implement, work reliably in conjunction with all certain low‐level machine instructions. this paper we present two novel hooking, which call Debugger‐aided Single...