Univerza na Primorskem Fakulteta za matematiko, naravoslovje in informacijske tehnologije
SI | EN

ponedeljek, 30. junij 2025 Una VULETIĆ: Computational Analysis and Visualization of Viewing Behavior: Interactive vs. Non-Interactive Documentary / Zala ŽUŽEK: Prediction of Eudaimonic and Hedonic Movie characteristics using User reviews

V ponedeljek, 30. junija 2025, bosta ob 16:00 uri izvedeni dve
predavanji v okviru PONEDELJKOVEGA SEMINARJA RAČUNALNIŠTVA IN INFORMATIKE
Oddelkov za Informacijske znanosti in tehnologije UP FAMNIT in UP IAM.

ČAS/PROSTOR: 30. junij 2025 ob 16.00 v FAMNIT-VP2.

1. predavanje:
============

------------------------------------------
PREDAVATELJICA: Una VULETIĆ
------------------------------------------

Una Vuletić is 2nd year Data Science master's student at UP FAMNIT, writing her master thesis under supervision of Associate Professors Matjaž Kljun and Klen Čopič Pucihar. She is also working at HICUP Lab at UP FAMNIT, as part of the pilot project GDI UP.

---------------------------------------------------------------------------------------------------------------------------------------------------------------
NASLOV: Computational Analysis and Visualization of Viewing Behavior: Interactive vs. Non-Interactive Documentary
---------------------------------------------------------------------------------------------------------------------------------------------------------------

POVZETEK:

In this seminar, we explore how interactivity in digital documentaries influences viewer experience, both visually and emotionally. While traditional documentaries rely on passive, linear storytelling, which can limit emotional engagement and attention, interactive documentaries allow viewers to shape their own experience while navigating through content. To investigate effect of interactivity, we analyzed eye-tracking and facial expression data from two participant groups who watched interactive and non-interactive versions of the iIsland documentary, which focuses on the lives of the last residents of Biševo, a small Croatian island.

Seminar bo potekal v angleškem jeziku.

=============================================================================================================

2. predavanje:
============

---------------------------------------
PREDAVATELJICA: Zala ŽUŽEK
---------------------------------------

Zala Žužek is currently completing her master’s studies in Data Science at UP FAMNIT. Her interests include data analytics, machine learning, and recommender systems, which she explores in depth through her master’s thesis on the prediction of psychological content in user-generated movie reviews. She currently works as a data analyst in the Data Intelligence and Delivery Assurance department, where she applies and deepens her expertise in data analytics, process automation, and data-driven assurance. Her work involves automating audit processes, improving data quality, and supporting digital transformation by integrating advanced analytics into assurance workflows. She is passionate about using data to enhance efficiency, uncover meaningful insights, and enable smarter decision-making in complex business environments.

-------------------------------------------------------------------------------------------------------------------------
NASLOV: Prediction of Eudaimonic and Hedonic Movie characteristics using User reviews
-------------------------------------------------------------------------------------------------------------------------

POVZETEK:

This thesis explores the use of machine learning to predict eudaimonic and hedonic movie characteristics based on user reviews. Using natural language processing (NLP) techniques, we extracted relevant features from full-text reviews scraped from an online movie database and trained several regression and classification models. The results indicate that user reviews contain sufficient emotional and semantic information that can predict user preferences more effectively than baseline models. These findings suggest that integrating text-based insights with traditional features can provide a deeper understanding of user behavior and improve the performance of recommendation systems.

Seminar bo potekal v angleškem jeziku.

 

Vabljeni.