This research aims to reduce dropout rates in higher education by developing a machine learning model to predict churn early, enabling timely interventions. The most important research findings of the thesis are summarized as follow: - The overall dropout rate at the University of Banja Luka, Bosnia and Herzegovina, was nearly half of enrolled students between the 2007/08 and 2018/19 academic years, with rates showing an increase over time. - Half of the student churn occurs within the first year of enrollment, with significantly higher rates in three-year study programs compared to four-year programs. - The results suggest that at least one-third of student churn could potentially be prevented, as it is attributable to institutional factors. - The findings indicate that it is possible to predict student churn at the earliest stages of education, even when using a challenging dataset with missing data and limited pre-academic, academic, and socioeconomic features. - The use of the Histogram-based Gradient Boosting Classifier (HGBC) in the educational data field, for the first time, resulted in an attrition classification accuracy of 75% at the beginning of the first academic year and 83% by the end, outperforming other widely used models. - To validate the effectiveness of HGBC, the model was tested under varying conditions and consistently demonstrated high performance. - The application of pre- and post-hoc interpretability techniques can enhance the credibility and reliability of machine learning models in predicting student discontinuation. - The most significant predictors of churn include gender (female), student cohort, accumulated ECTS credits, scholarship status, age at enrollment, the number of successfully passed courses, program duration, and whether the student attended a Gymnasium. - Certain variables, such as gender (female) and student cohort, consistently maintained a high level of significance across the models over time. - Models with lower performance and quality exhibited a similar ranking of feature importance to those of the best-performing models.

to provide

Aim: Coronary artery disease is a leading cause of death in Croatia, and percutaneous coronary intervention (PCI) has emerged as a significant advancement in its treatment. The facilitation of PCI is performed by an interventional team consisting of an interventional cardiologist, a nurse, and a radiologic technologist. In order to achieve precise and efficient intervention outcomes, the radiologic technologist’s role in operating the fluoroscopy is crucial. This study’s aim was to investigate the attitudes of undergraduate professional radiologic technology students at the Faculty of Health Studies in Rijeka regarding the role of the radiologic technologist within the specialist team during PCI and the adequacy of competencies acquired during university education. Materials and methods : The research included students from all three years of regular undergraduate professional studies in radiologic technology, representing diverse demographic characteristics. Data were collected online through a questionnaire developed in Google Forms and analyzed using the Statistica software program. Results : The majority of respondents view the radiologic technologist as essential in the medical team during PCI, with effective communication being emphasized as a crucial element for successful teamwork. However, respondents express a belief that their previous education has not equipped them with adequate knowledge and skills to function com-petently within a PCI team, indicating a perceived need for additional education and professional training. Conclusion : The study shows that radiologic technology students comprehend the significance of radiologic technologists in the interventional team for PCI procedures. However, they express concerns regarding the insufficiency of knowledge and skills acquired during their studies for proficient work in interventional cardiology laboratories.

Motivation: Quantitative MRI (qMRI) offers sensitive and specific measures to study age-related microstructural changes in the brain. However, models assessing age trajectories in qMRI brain properties are often incomparable among centers. Goal(s): Develop normative models reflecting aging trajectories and assess the impact of bi-centric, non-fully matched protocols in brain aging studies. Approach: Investigating age trajectories in cortical regions using polynomial regression models, focusing on quantitative R1, R2*, and susceptibility mapping (QSM). Results: We validated data harmonization by observing the impact on normative trajectories using bicentric data, where we noted significantly different maturation and aging inflections for R1 and R2* trajectories across cortical regions. Impact: This bi-centric, multi-parameter qMRI study investigates age-dependent variations across cortical regions, offering a valuable reference for subsequent qMRI aging research and emphasizing age effects on the cortical surface.

Motivation: Progression independent of relapse activity (PIRA) is the most frequent manifestation of disability accumulation in multiple sclerosis (MS), but the mechanisms leading to PIRA are currently unknown. Goal(s): To investigate the link between PIRA and white matter degeneration in people with MS. Approach: To compare the integrity of normal-appearing white matter (NAWM) between patients with MS who experienced PIRA versus stable patients using diffusion tensor imaging (DTI) measures from a clinical-compatible protocol. Results: Patients with PIRA exhibited significant differences in DTI-derived measures compared to stable patients: reduced fractional anisotropy and increased mean and radial diffusivity in NAWM. Impact: This study sheds light on the relationship between progression independent of relapse activity (PIRA) and white matter degeneration in people with multiple sclerosis. The results have important implications for understanding the mechanisms of disability progression in relapsing-remitting multiple sclerosis.

For predicting cardiovascular diseases, mathematical modelling of the cardiovascular system has been proven to be a powerful asset. The governing idea is to analyse it through compartments as multiple connected subsystems with inputs and outputs. In this paper, models were identified for four subsystems of input-output sequence (left ventricle - left atrium - ascending aorta - descending aorta - left common carotid artery) by modelling frequency response. The data set used for model identification consisted of blood pressure during four consecutive heart contractions of four circulatory segments from clinical trials performed on a pig. The goal is to discover a linear model with a non-integer order that succinctly represents the process, outperforming high-order autoregressive exogenous input (ARX) integer models. This model identification occurs non-parametrically, aiming to achieve the best smooth fit in the frequency domain by minimizing the difference between real measurements and model predictions using the particle swarm optimization (PSO) algorithm.