https://rfos.fon.bg.ac.rs/handle/123456789/3 2026-11-09T04:33:43Z https://rfos.fon.bg.ac.rs/handle/123456789/3294 Title: Classifying Expected Benefits of Collective Organization: A Rule-Based Approach for Non-Wood Forest Product Collectors in Serbia Authors: Nedeljković, Jelena; Đukić, Daliborka; Milošević, Pavle; Nonić, Dragan; Živojinović, Ivana 2026-10-23T00:00:00Z https://rfos.fon.bg.ac.rs/handle/123456789/3293 Title: From rookie to contract extension: Modeling NBA player career trajectories with machine learning Authors: Milošević, Pavle; Milošević, Dejan; Milenković, Ivan; Dragović, Ivana; Rakićević, Aleksandar; Zukanović, Milica Abstract: This paper investigates machine learning-based prediction of players’ long-term career trajectories at different career stages. Specifically, we predict whether a player is likely to achieve Elite Player, High-Impact Starter or Rotation Player status based on data from their first one, three, and five NBA seasons. The target classes are expert-defined, subjective categories intended to capture meaningful differences in players’ career success and on-court impact. We collected, analyzed, and preprocessed a dataset of 942 players, described by 60 attributes representing their performance metrics and physical characteristics. We compare the classification performance of Multilayer Perceptron, Random Forest, and XGBoost, along with Logistic Regression as a benchmark. Furthermore, we conducted two experiments: one that takes into account the effect of NBA era player was active and another that neglects it. Finally, we identify the most important attributes contributing to classification outcome. Random Forest achieves the best performance after the rookie season (ROCAUC = 0.671 ± 0.039), indicating that career outcomes can be predicted only to a limited extent from rookie-season data. Performance improves substantially after three seasons (0.793 ± 0.025) and five seasons (0.859 ± 0.023), suggesting that additional performance history provides much stronger signals of long-term career success. PER is identified as the most important performance indicator. 2026-08-27T00:00:00Z https://rfos.fon.bg.ac.rs/handle/123456789/3292 Title: Profiling Cognitive Offloading in LLM-Mediated Synthesis Writing: Volume vs. Content Authors: Poquet, Oleksandra; Nanduri, Mani Shankar; Loyer, Maria Ximena Salinas; Stadler, Matthias; Sailer, Michael; Jovanović, Jelena Abstract: This study compares two approaches to profiling how learners offload cognitive activity to LLMs during a synthesis writing task. Drawing on Salomon’s distributed cognition and the Kintsch and van Dijk model of text comprehension, the study operationalises offloading to an LLM in two ways: as a volume of LLM use and as content of what is offloaded, both along with prior knowledge. Data from 97 university students using a ChatGPT-based interface were analysed using k-means clustering. To capture the content of offloading, their prompts were interpreted as to who performs the activity (active/passive) and at what level of comprehension (local/global). Volume-based profiling (k=4) differentiated learners primarily by prior knowledge, with volume negatively associated with essay authorship. Content-based profiling (k=5) revealed qualitatively distinct patterns of offloading, from vocabulary clarification to active direction of structuring and generation to passive delegation of comprehension at both levels. These patterns reflect different fragmentation of the cognitive process, with significant differences in learning strategies, behavioural markers, and essay authorship. Combining volume and content of offloading could improve future analyses on how LLM use redistributes cognitive activity and its effects on learners. 2026-10-07T00:00:00Z https://rfos.fon.bg.ac.rs/handle/123456789/3291 Title: Understanding How Learning Design Co-authoring Unfolds in Online Platforms Authors: Rodríguez-Triana, María Jesús; Jovanović, Jelena; Prieto, Luis Pablo; Poquet, Oleksandra; Ley, Tobias; Gillet, Denis Abstract: Online platforms for learning design (LD) co-authoring offer teachers opportunities to engage with pedagogical practice in a community. Supporting co-design at scale requires understanding its temporal unfolding, often relying solely on trace data. Yet, most studies of learning (co-)design focus on group-level analyses of the quality and composition of the finished LD product, rather than the heterogeneous unfolding of its process. This study analyses patterns of trace data from the widely used platform, Graasp, using the Knowledge Appropriation Model (KAM) to examine how such co-design processes evolve over time. Clustering 1,050 designs, we identify LD co-authoring profiles based on features reflecting KAM maturation phases as well as process predictability and stability indicators extracted using recurrence quantification analysis. These profiles are then analysed based on a critical phase transition in the teachers’ learning process: the LD’s classroom implementation. Our results highlight effort, regularity, and co-creation (multiple authors contributing content) as key to successful LD co-authoring that transitions to implementation. The study also underscores challenges in applying theoretical constructs to analyse trace data and limitations of such data in capturing complex, blended social practices. 2026-10-07T00:00:00Z