Natural language processing applications in social network analysis: a data mining approach

Abstract Social media texts, such as tweets and user profiles, reflect the opinions, emotions, preferences, and behaviors of social media users, as well as the structure and dynamics of their social networks. Analyzing social media texts can provide valuable insights and applications for various domains and stakeholders, such as marketing, politics, healthcare, and education. However, analyzing social media texts poses significant challenges, as they are often noisy, informal, unstructured, and heterogeneous. Moreover, social media texts are not only influenced by the individual characteristics and contexts of the users but also by the complex and evolving interactions and relationships among them. Therefore, traditional methods and techniques for natural language processing (NLP), social network analysis (SNA), and data mining (DM) may not be sufficient or effective for social media text analysis. In this paper, we propose a data mining approach that integrates and leverages NLP, SNA, and DM methods and techniques for social media text analysis. We demonstrate the feasibility and effectiveness of our approach on a large and comprehensive data set of tweets and user profiles related to the domain of consumer spending. We show that our approach can provide comprehensive and effective solutions and applications for various domains and stakeholders, such as marketing, and education. We also compare and contrast our results and outcomes with the existing works, and show the advantages and improvements of our approach over the existing works.