Design science research (DSR) is one of important research paradigms in information systems (IS) that focus on addressing business problems by building and implementing design artifacts. Recently, predictive analytics has become one major stream of DSR thanks to the improvement of computational power and methods and the increase in available datasets. While predictive analytics studies have been much effective in producing practical implications, often, they have been criticized for lacking theoretical contributions. As a response to this criticism, in this dissertation, I present two different approaches for theory-driven predictive analytics in the context of social media issues: theory-driven feature engineering and theory-driven model designing. Social media has been a mixed blessing to our society presenting both merits (e.g., easy access to information, social support, etc.) and demerits (e.g., fake news, hate speech, etc.). Lately, predictive analytics has been largely adopted as an effective methodological approach to create solutions to the issues of social media. Through the collection of three essays included in this dissertation, I demonstrate how the two theory-driven approaches for predictive analytics can be implemented in real-world settings. In detail, the first essay proposes a novel method to represent online text as signed networks based on the structural balance theory, which are fed as an important information source into a deep learning model that identifies false information. The second essay develops an automated fake news detection model that takes into account deceptive intention behind news publishers. The third essay leverages previous studies on personality traits and hate behavior to develop a deep learning model for identifying online hate speech. Rigorous evaluation reveals that the set of predictive models proposed in this dissertation not only address important social challenges but also broadly contributes to the literature in predictive analytics, design science, and IS research. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com.bibliotheek.ehb.be/en-US/products/dissertations/individuals.shtml.]