Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.736779
Title: Analysing people’s sentiment and emotional reaction towards online videos
Author: Mulholland, Eleanor
ISNI:       0000 0004 6500 853X
Awarding Body: Ulster University
Current Institution: Ulster University
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Abstract:
The purpose of this research is to evaluate the quality of online video content. Watching video content online has become one of the most popular activities on the Internet as audiences continue to grow rapidly around the globe where more people than ever access the Internet. This leads to a global demand for more video content and an ever increasing number of video content producers. As online audiences grow and the number of video content producers increases a challenge is the identification of high quality video content amongst billions of hours of video. Evaluating video quality and identifying video content that will entertain and emotionally engage the user is a complex task. Evaluating high quality video content has commonly been addressed with video recommender systems, which rank videos for comparison. Sentiment analysis, which identifies positive and negative emotions in text has been applied to text movie reviews to improve movie recommendations. Affective recommender systems utilise the emotional state of the user to improve recommendations. Gamification, which utilises game techniques in non-game scenarios, is a recognised method of encouraging user participation and for monitoring online communities. This research focuses on the use of sentiment analysis, emotion detection and gamification in an emotion-centred model for the evaluation of online video content. The emotion-centred model combines the Unifying framework (Tkalcic et al. 2011) and the Emotion Imbued Choice model (Lerner et al. 2015). The emotion-centred model is implemented in a software system called 360-MAM-Affect. 360-MAM-Affect's sentiment analysis module automatically evaluates the quality of online videos by analysing people's text comments on them. 360- MAM-Affect's emotion detection module obtains people's emotion feedback on videos together with data on their emotional state including explicit mood feedback, implicit mood feedback, and personality and preferences. 360-Gamify, a gamification module, uses gamification techniques to encourage the user to proactively provide feedback and engage with 360-MAM-Affect. This thesis investigates two hypotheses: (1) Video-Sentiment Hypothesis (Hl): Sentiment analysis can enhance the quality evaluation o f online videos and (2) Video-Emotional-Reaction Hypothesis (H2): Users can be monitored with emotion detection and gamification during video viewing to identify their current emotional state and emotional reaction to video content. Five experiments are conducted on YouTube videos with 360-MAM-Affect in order to investigate Hl and H2, with 2 of these experiments applied to 1,433 YouTube videos, and one experiment involving 22 human participants interacting with 200 of those YouTube videos over a two-week period. The results from both experiments provide evidence for hypotheses Hl and H2. Future work includes experimentation with a larger number of participants and incorporating 360-MAM-Affect within a video recommender system.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.736779  DOI: Not available
Share: