Repository of Faculty of Science
Not a member yet
1747 research outputs found
Sort by
Classification of book genres using natural language processing techniques
No full textIstražena je klasifikacija žanrova knjiga koristeći tehnike obrade prirodnog jezika. Cilj je bio razviti modele koji precizno predviđaju žanrove knjiga na temelju tekstualnih opisa, primjenom metoda poput TF-IDF-a, Word2Vec-a i algoritama kao što su logistička regresija i slučajna šuma. Rezultati pokazuju da kombinacija naprednih vektorizacija i klasifikacijskih algoritama pruža visoku točnost u klasifikaciji višestrukih žanrova. Identificirani su izazovi vezani uz neuravnoteženost podataka te su predložene smjernice za daljnja istraživanja.The classification of book genres was explored using natural language processing techniques. The goal was to develop models that accurately predict book genres based on textual descriptions, employing methods such as TF-IDF, Word2Vec, and algorithms like logistic regression and Random Forest. The results indicate that the combination of advanced vectorization techniques and classification algorithms provides high accuracy in multi-label genre classification. Challenges related to data imbalance were identified, and guidelines for further research were proposed
Performance comparison of YOLOv5 and YOLOv8 algorithms in classification tasks
No full textUmjetna inteligencija predstavlja široko interdisciplinarno područje informatike koje stvara sustave sposobne demonstrirati ljudske vještine poput percepcije, učenja i rješavanja problema. Rad uvodi u koncepte ključne za razvoj AI sustava, kao strojno učenje i duboko učenje. Posebno se ističe YOLO algoritam dubokog učenja za detekciju objekata, razvijen radi brze i precizne analize slika. Pregledavajući arhitekturu varijanti v5 i v8 YOLO modela, ističe se njegova sposobnost prilagodbe različitim zahtjevima te podrška za vlastiti skup podataka. Analiza metrika poput srednje prosječne preciznosti i F1 pruža uvid u evaluaciju performansi modela. Treniranje modela, ključno za postizanje visokih performansi, ističe važnost kvalitetnih podataka i finog podešavanja (engl. fine-tuning) za specifične potrebe aplikacije. Detekcija objekata kroz klasifikaciju i regresiju objašnjava kako YOLO model brzo i precizno detektira objekte u slikama. Kroz primjere primjene, rad ilustrira kako pripremiti skup podataka i trenirati na njemu YOLO model, te uspoređuje različite inačice algoritma preko metrika i performansi.Artificial intelligence represents a broad interdisciplinary field of computer science that creates systems capable of demonstrating human-like skills such as perception, learning and problem solving. The paper introduces key concepts essential for the development of AI systems, such as machine learning and deep learning, where the YOLO algorithm for object detection is notable. By reviewing the architecture of the YOLO v5 and v8 model variants, the focus is on its adaptability to various requirements and support for custom datasets. The analysis of metrics such as mean average precision and F1 provides insight into the evaluation of model performance. Model training, crucial for achieving high performance, emphasizes the importance of quality data and fine-tuning for specific application needs. Object detection through classification and regression explains how YOLO model quickly and accurately detects objects in images. Through practical application, the paper illustrates how to prepare a dataset and train the YOLO model on it, comparing different versions of the algorithm based on metrics and performance
The role of retinoids in skin health: mechanisms, benefits and future directions
No full textRetinoidi, sintetski ili prirodni derivati vitamina A, koriste se u tretiranju promijenjenog izgleda kože, koji nastaje zbog intrinzičnog i ekstrinzičnog starenje ili različitih kožnih oboljenja poput akni. Kako bi došli do ciljnih stanica, moraju proći pretvorbe iz jednog oblika u drugi, a zatim vezani za stanične retinoid-vezujuće proteine (CRABP) doći do nuklearnih receptora. Retinol, jedan od derivata vitamina A, zbog svoje lipofilne prirode prolazi do epidermisa i dermisa. Povećava debljinu epidermisa, proliferaciju keratinocita, broj dermalnih proteina izvanstaničnog matriksa i broj krvnih žila te tako poboljšava nepravilan i oštećen izgled kože. Također, retinoidi koji su se pokazali učinkovitima za liječenje akni su trifaroten i oralni izotretinoin. Ponekad primjena retinoida, osim pozitivnih utjecaja, uzrokuje nuspojave. Neke od njih su: svrbež, peckanje, eritem, suhoća i ljuskanje. Kako bi izbjegli negativne posljedice korištenja retinoida znanstvenici su u potrazi za novim unaprijeđenim formulacijama retinoida kao što su nanočestice, liposomi, mikroemulzije i mikrospužve.Retinoids, synthetic or natural derivatives of vitamin A, are used to address the changed appearance of the skin, which occurs due to intrinsic and extrinsic aging or various skin conditions such as acne. To reach the target cells, they must undergo conversion from one form to another, and then bound to cellular retinoid binding protein (CRABP) reach nuclear receptors. Retinol, one of the derivatives of vitamin A, penetrates the epidermis and dermis due to its lipophilic nature. It increases the thickness of the epidermis, the proliferation of keratinocytes, the number of dermal extracellular matrix proteins and the number of blood vessels, thus improving the irregular and damaged appearance of the skin. Also, retinoids that have shown a positive impact on the appearance of skin with acne are trifarotene and oral isotretinoin. Sometimes the use of retinoids, in addition to positive effects, causes side effects. Some of them are itching, burning, erythema, dryness and scaling. To avoid the negative consequences of using retinoids, scientists are looking for new, more stable formulations of retinoids such as nanoparticles, liposomes, microemulsions and microsponges
Students' Preferences and Understanding of Relational and NoSQL Databases
No full textOvaj diplomski rad istražuje stavove studenata Prirodoslovno-matematičkog fakulteta u Splitu prema različitim vrstama baza podataka, s posebnim fokusom na relacijske i nerelacijske baze podataka. Cilj istraživanja bio je analizirati preferencije studenata, njihovo razumijevanje rada s bazama podataka te percepciju primjene tih baza u budućem profesionalnom životu. Rad uključuje pregled ključnih pojmova vezanih uz baze podataka i modeliranje podataka, te opisuje ulogu alata za učenje, poput LearnER, koji studentima pomažu u svladavanju vještina modeliranja. Kroz istraživanje su analizirani rezultati anketa i testova provedenih na studentima različitih smjerova i razina studija, a korištene su i statističke metode za procjenu značajnosti razlika među skupinama. Rezultati pokazuju da nema značajnih razlika u preferencijama prema vrstama baza podataka među studentima različitih studijskih smjerova i razina studija, ali postoje razlike u razini razumijevanja modeliranja podataka. Rad zaključuje da studenti pokazuju umjeren interes i samopouzdanje u radu s bazama podataka te ističe važnost prilagođavanja nastavnih metoda kako bi se unaprijedilo njihovo razumijevanje i praktične vještine u ovom području.This thesis explores the attitudes of students at the Faculty of Science, University of Split, towards different types of databases, with a particular focus on relational and non-relational databases. The aim of the research was to analyze students' preferences, their understanding of working with databases, and their perception of applying these databases in their future professional lives. The thesis includes an overview of key concepts related to databases and data modeling and describes the role of learning tools, such as LearnER, which help students acquire modeling skills. The research involved analyzing survey and test results conducted on students of different study programs and levels, using statistical methods to assess the significance of differences between groups. The results indicate no significant differences in preferences for types of databases among students of different study programs and levels, but there are differences in the level of understanding of data modeling. The thesis concludes that students show moderate interest and confidence in working with databases, and emphasizes the importance of adapting teaching methods to improve their understanding and practical skills in this field
Comparison of belt and chain transmission
No full textRemenski i lančani prijenosnici se vrlo često koriste za prijenos snage između vratila. Razlikuju se po svojoj konstrukciji te načinu prijenosa. U radu su prikazane vrste remenskog prijenosa koji se razlikuju ovisno o obliku remena koji koriste. Remenski prijenosnici s plosnatim i klinastim remenom prenose snagu trenjem, dok remenski prijenosnici sa zupčastim remenom to čine oblikom. Lančani prijenosnici koriste lanac za prijenos snage i gibanja između zupčanika na različitim vratilima. Upotrebljavaju se razne vrste lanaca, koji se biraju ovisno o opterećenju i obodnoj brzini. U radu osim karakteristika prijenosnika prikazani su materijali koji se koriste za njihovu izradu, navedena su područja njihove primjene te prednosti i nedostatci.Belt and chain drives are very often used to transmit power between shafts. They differ in their design and the type of transmission. The paper presents the types of belt drives that differ depending on the shape of the belt used. Belt drives with flat and V-belts transmit power through friction, while belt drives with toothed belts do so through their shape. In chain drives, a chain is used to transmit power and motion between sprockets on different shafts. Various types of chain are used, which are selected depending on the load and circumferential speed. In addition to the properties of belt and chain drives, the materials used to manufacture them, their areas of application and their advantages and disadvantages are also presented
The role of vitamin E and vitamin E acetate on monolipid and bilipid membranes
No full textVitamin E esencijalni je nutrijent s ključnom ulogom u zaštiti stanica od oksidativnog stresa i održavanju zdravlja. Obzirom da je riječ o supstanci koju ljudsko tijelo ne može samostalno sintetizirati, sve je češća industrijska proizvodnja kojom se dobiva esterizirani oblik, vitamin E acetat. Vitamin E acetat je stabilan i manje reaktivan oblik vitamina E. Iako zbog svoje otpornosti na oksidaciju, acetatna forma pruža brojne prednosti te se stoga često koristi u prehrambenim dodacima i kozmetici, važno je biti svjestan potencijalnih rizika, osobito kod inhalacije koja je u posljednje vrijeme sve aktualnija medu mlađom populacijom. Tema ovog rada je istaknuti izrazito pozitivnu funkciju vitamina E u borbi protiv slobodnih radikala, ali obraditi i njegov potencijalno negativni utjecaj na ljudski organizam ukoliko dođe do monolipidne membrane plućnog surfaktanta što za posljedicu može imati razvoj potencijalno terminalne bolesti “EVALI” (engl. “E-cigarette or Vaping Use-Associated Lung Injury“) te na taj način predstaviti jedan veoma aktualan problem.Vitamin E is an essential nutrient that plays a key role in protecting cells from oxidative stress and maintaining health. As it is a substance that the human body cannot synthesize itself, it is increasingly being produced industrially to obtain an esterified form, vitamin E acetate. Vitamin E acetate is a stable and less reactive form of vitamin E. Although the acetate form offers many advantages due to its resistance to oxidation and is therefore often used in food supplements and cosmetics, one must be aware of the potential risks, especially with inhalation, which has recently become increasingly common in the younger population. This paper aims to highlight the extremely positive function of vitamin E in the fight against free radicals, but also its potentially negative effects on the human body when the monolipid membrane of the pulmonary surfactant is attacked, which can lead to the development of a potentially fatal disease "EVALI" („E-cigarette or Vaping Use-Associated Lung Injury“)
Application and Comparison of LLMs in Code Evaluation
No full textOvo istraživanje za cilj ima istražiti primjenu i usporedbu velikih jezičnih modela u evaluaciji procjene koda. Usredotočeno je na tri modela: OpenAI-jev GPT-4, Code Llama 7B Instruct i fino podešenu verziju CodeT5+. Modeli se ocjenjuju na temelju njihove sposobnosti procjene koda u različitim kategorijama, uključujući matematičke i algoritamske zadatke, strukture podataka i web programiranje. Uspoređujući izvedbu ovih modela, nastoje se identificirati njihove prednosti i slabosti u pružanju točnih i korisnih povratnih informacija za poboljšanje kvalitete koda. Ovo istraživanje naglašava potencijal velikih jezičnih u pojednostavljivanju procesa ocjenjivanja programskog koda, a istovremeno prepoznaje potrebu za daljnjim napretkom u njihovom finom podešavanju i primjeni.This research aims to investigate the application and comparison of large language models in the evaluation of code estimation. It focuses on three models: OpenAI's GPT-4, Code Llama 7B Instruct, and a fine-tuned version of CodeT5+. Models are evaluated based on their ability to evaluate code in a variety of categories, including mathematical and algorithmic tasks, data structures, and web programming. By comparing the performance of these models, we seek to identify their strengths and weaknesses in providing accurate and useful feedback to improve code quality. This research highlights the potential of large languages in simplifying the process of evaluating software code, while recognizing the need for further progress in their fine-tuning and application
Analiza molekularne dinamike PQQ-GDH proteina u interakciji s ugljikovom nanocjevčicom
No full textInterest for enzymatic biofuel cells grows in the past years, due to their biological components which counteract noble-metal and alloy-based fuel cells. Soluble glucose dehydrogenase, or PQQ-GDH enzyme, showed exciting possibility for application in biosensing and bioenergetics, due to their high catalytic activity which is oxygen independent. Analysis of molecular dynamics simulations of the monomer of protein PQQ-GDH in complex with short and infinitely long carbon nanotube (CNT) has been carried out, using methods and tools of VMD program. To examine the stability and the strength of protein nanotube interactions, distances between certain residues and CNT wall were calculated. To examine protein stability and function preservation, root-mean-square deviation and radius of gyration of monomers with and without nanotube were computed. Finally, considering results from simulation analysis, a theoretical prototype of the enzymatic biosensor comprising PQQ-GDH enzyme in complex with a short CNT was constructed.Interes za enzimatske biogorivne ćelije raste proteklih godina, zbog njihovih bioloških komponenti koje se suprotstavljaju gorivnim ćelijama na bazi plemenitih metala i legura. Topljiva glukoza dehidrogenaza, ili enzim PQQ-GDH, pokazao je obećavajuće pokušaje primjene u biosenzorima i bioenergetici, zbog svoje visoke katalitičke aktivnosti neovisne o kisiku. Koristeći metode i alate programa VMD, provedena je analiza simulacija molekularne dinamike monomera proteina PQQ-GDH u kompleksu s kratkom i beskonačno dugom ugljikovom nanocjevčicom (CNT). U svrhu ispitivanja stabilnosti i snage interakcija između monomera i nanocjevčice, izračunate su udaljenosti između određenih aminokiselinskih ostataka i stijenke CNT-a. U svrhu ispitivanja stabilnosti i očuvanja funkcije proteina, izračunate su standardne devijacije i radijusi giracije monomera sa i bez nanocjevčice. Na kraju, uzimajući u obzir rezultate analize simulacija, konstruiran je teorijski prototip enzimatskog biosenzora koji sadrži PQQ-GDH enzim u kompleksu s kratkom CNT
Adapting Digital Tools for Children with Autism Spectrum Disorder: The Role of Educational Platforms and Participatory Design
No full textSve veći broj djece s poremećajima iz spektra autizma (ASD) zahtijeva prilagodbu korisničkog sučelja zbog specifičnih teškoća s kojima se suočavaju. Ovaj rad istražuje ulogu digitalnih alata u unapređenju obrazovnih i društvenih iskustava djece s ASD-om. U okviru istraživanja, provedena je identifikacija relevantnih digitalnih alata i aplikacija koje pomažu u razvoju socijalnih i obrazovnih kompetencija djece. Detaljno su analizirani obrazovni alati poput TaLNA i TeachTown te okruženje ECHOES, koje osim obrazovne funkcije potiče razvoj društvenih vještina kod djece s ASD-om. Također je istražen utjecaj nekoliko tablet aplikacija na razvoj socijalnih kompetencija djece s ASD-om. Završni dio rada fokusira se na primjenu participativnog dizajna, s ciljem prilagodbe digitalnih alata djetetovim individualnim potrebama, preferencijama i sposobnostima, što je ključ za postizanje učinkovitije interakcije s tehnologijom.There are more and more children with autism spectrum disorder who, due to their difficulties, require the adaptation of the user interface to their needs. This paper explores the role of digital tools in enhancing the educational and social experiences of children with ASD. The educational tools TaLNA, TeachTown and the ECHOES environment are explained in more detail. ECHOES, in addition to its educational role, serves to encourage the social skills of children with ASD. Several tablet applications and their influence on the development of children's social competences are also presented. At the end of the paper, the process of participatory design for children with ASD is presented with the aim of adapting digital tools to the child's preferences and capabilities
Dynamics of circular movement
No full textKružno gibanje podrazumijeva vrtnju tijela oko neke osi. Kako bismo to gibanje opisali i odredili njegov uzrok, koristimo se linearnim i kutnim kinematičkim i dinamičkim veličinama. Cilj ovog rada je istražiti fizikalne veličine o kojima ovisi moment tromosti tijela te proučiti i odrediti utjecaj momenta tromosti na kinematičke i dinamičke veličine tijekom kružnog gibanja. Istraživanje je izvedeno na novo-konstruiranom eksperimentalnom uređaju, a večina podataka prikupljena je kamerom budući da smo snimali gibanje koje je uzrokovalo rotacijsko gibanje tijela specifičnog oblika. Glavnina obrade podataka učinjena je koristeći softver za video analizu i modeliranje u fizici Tracker. Metodom najmanjih kvadrata određen je moment tromosti tijela nepravilnog oblika te je eksperimentalno pokazana proporcionalanosti s masom tijela i kvadratom radijusa na kojem se masa nalazi. Dodatno, na temelju ovog rada izrađena je i praktikumska vježba.A circular motion implies a rotation of the body around an axis. To describe this motion and determine its cause, we use linear and angular kinematic and dynamic quantities. The aim of this work is to investigate the physical quantities on which the body’s moment of inertia depends and to investigate and determine the influence of the moment of inertia on the kinematic and dynamic quantities in a circular motion. The investigation was carried out on a newly constructed experimental device. Most of the data was collected using a camera, as we recorded the motion caused by the rotational movement of a particular irregularly shaped body. Majoritiy of the data processing was done with the video analysis and physics modeling software Tracker. The moment of inertia of a body of a specific shape was determined using the method of least squares, and the proportionality with the mass of the body and the square of the radius in which the mass is located was experimentally proven. In addition, a student laboratory exercise was created on the basis of this work