TU home | T3 | Information Technology Dissertation Defense Announcement for Youngsub Han

Information Technology Dissertation Defense Announcement for Youngsub Han

The Information Technology Doctor of Science program invites the university community to a dissertation defense for Youngsub Han on April 19, 2017 at 11:00 am in York Road room 459.

Dissertation Title: A Study of Aspect-Based Sentiment Analysis in Social Media

 

Abstract:

In recent years, diverse social media platforms, such as Facebook, YouTube, Instagram, Snapchat, and Twitter, have rapidly grown in size and influence. Various industries, including the media and advertising industries, have made significant efforts to adapt to building competitive advantages using social media in the marketplace. Therefore, it became imperative for marketing and advertising professionals to better understand the complex behaviors and minds of consumers using data-mining techniques, which help to handle massive amounts of social media data.

In this research, we analyzed Twitter to discover characteristics of social media. This study is intended to address these topics to build a better understanding of Twitter usages. Even if there are many important theories and frameworks in media studies, we were not able to single out one specific theoretical framework for this study. Therefore, by using the active audience concept, and relying on marketing literature, we chose a grounded theory approach and presented research questions for in-depth understanding of Twitter usage in order to detect any patterns that consumers might show.

In addition, we proposed a sentiment analysis method and a lexicon building method using the morphological sentence pattern (MSP) model. These methods aim to observe and summarize people’s opinions or emotional states from textual data. Despite the demands of sentiment analysis methods for analyzing social media data, fundamental challenges remain because user-generated online textual data is unstructured, unlabeled, and noisy. Finally, we proposed a method of extracting movie genre similarity for movie recommendation using our proposed methods.