Advanced user experience analysis in digital products using artificial intelligence sentiment analysis techniquesDunoska, Emilija (Avtor) Šumak, Boštjan (Mentor) Brdnik, Saša (Komentor) user experiencesentiment analysisfacial expression recognitionself-reported affectdigital productsArtificial intelligence is increasingly being used for the detection of users' emotions, yet limited evidence exists comparing AI-based methods with established self-report measures. This master's thesis addresses this gap by examining the performance of video-based and text-based sentiment analysis methods against a validated self-report measure of affect. A controlled experiment was conducted with 27 students completing tasks on two digital platforms, during which facial expressions and written reflections were collected. Alignment with self-reports was observed for the text-based method in nearly three-quarters of cases, while positive experiences were consistently misclassified as negative by the video-based tool, suggesting that a more trustworthy overview of users’ emotional experience is currently provided by textual analysis. Multimodal approaches combining multiple input data sources are recommended for reliable evaluation of user experience in digital products.[E. Dunoska]20262026-06-16 08:07:59Magistrsko delo/naloga98475UDK: 159.942:004.8(043.2)COBISS_ID: 283555843sl