20.500.12556/DKUM-95743 Empirična primerjava znanstvenih objav s semantičnimi vektorji magistrsko delo Empirical comparison of scientific publications with semantic vectors V magistrski nalogi obravnavamo možnost uporabe semantičnih vektorskih reprezentacij kot orodja za primerjavo vsebine znanstvenih objav in spremljanje sprememb v raziskovalnih tematikah skozi čas. Menimo, da lahko z vektorizacijo povzetkov in merami kvantitativne razdalje zanesljivo sledimo podobnostim in razlikam v znanstveni produkciji. Posebej smo vključili tudi obdobje COVID-19, saj smo predvidevali, da bodo takratne vsebinske spremembe dober preizkus učinkovitosti metode. V raziskavi smo iz baze Scopus pridobili objave izbranih znanstvenih revij za obdobje 2017–2022, povzetke pretvorili v vektorje s prednaučenim SBERT modelom “all-MiniLM-L6-v2” ter zgradili predstavitve revij po letih. Za merjenje razdalj med vektorji smo uporabili kosinusno razdaljo, hipoteze pa preverili z Wilcoxonovim testom ter podprli rezultate z vizualizacijami in modeliranjem tematik z BERTopic. Rezultati potrjujejo, da obstajajo izrazite razlike med revijami, saj se vektorski centri različnih revij v prostoru jasno ločijo, medtem ko so objave znotraj posamezne revije bolj povezane. Prav tako smo zaznali tematske premike med zaporednimi leti iste revije, kar kaže na postopno, a zaznavno evolucijo raziskovalnih vsebin. Analiza obdobja pred in po letu 2020 je pokazala večje spremembe, vendar teh razlik ni mogoče neposredno pripisati COVID-19. Raziskava tako potrjuje, da je semantična vektorizacija skupaj s kosinusno razdaljo učinkovito orodje za empirično primerjavo vsebine znanstvenih objav. This master’s thesis explores the use of semantic vector representations to compare scientific publications and track shifts in research topics over time. We argue that vectorizing abstracts and applying quantitative distance measures can reliably capture similarities and differences in scientific production. The COVID-19 period was included as a potential test case for the method’s effectiveness. Publications from selected journals indexed in Scopus between 2017 and 2022 were collected. Abstracts were transformed into vectors using the pre-trained SBERT model “all-MiniLM-L6-v2”, and yearly journal representations were constructed. Cosine distance was applied to measure differences between vectors, while hypotheses were tested with the Wilcoxon signed-rank test. Results were further supported with visualizations and topic modeling using BERTopic. The findings confirm clear distinctions between journals, as vector centers of different journals are clearly separated, while publications within the same journal cluster more closely. Thematic shifts were also observed across consecutive years of the same journal, reflecting gradual but noticeable evolution of content. Analysis of the period before and after 2020 showed larger changes, though not directly attributable to COVID-19. Overall, the study demonstrates that semantic vectorization with cosine distance is an effective tool for empirical comparison of scientific publications. semantična vektorizacija znanstvene objave podobnost vsebine SBERT kosinusna razdalja semantic vectorization scientific publications content similarity SBERT cosine distance true true false [B. J. Javornik] Slovenski jezik Angleški jezik Magistrsko delo/naloga 2025-10-17 15:07:24 2025-12-22 09:39:25 2026-02-04 03:11:05 0000-00-00 00:00:00 2025 0 Maribor 0 Maribor 1 spletni vir (1 datoteka PDF (XII, 58 str.)) 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.4'414(043.2) 267099395 MAG_Javornik_Bojan_Jan_2025.pdf MAG_Javornik_Bojan_Jan_2025.pdf 1 5389392145C660BC0FE64EC05D5ACE41 913bac50d62574a459f2834a365ca36aa0948a7b66a53ca3238957d095229ea3 572450cb-ab71-11f0-a0a8-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=199993 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0