An Ontology-Driven Machine Learning Applications for Public Policy Analysis from Social Media Data: A Systematic Literature Review

Abstract

Admas Abtew, Dawit Demissie and Kula kekeba

With the widespread use of social media platforms for various purposes, including political discussions, public opinion on policy issues can be inferred from social media data. Ontology-driven machine learning techniques have been applied to extract relevant information from such data for public policy analysis. In this systematic literature review (SLR), we aim to identify the current state-of-the-art ontology-driven machine learning applications for public policy analysis from social media data. We conducted a comprehensive search for relevant studies published in peer-reviewed journals and conference proceedings up to May 2023. After screening and selection, we included 35 studies in the final review. Our review revealed that ontology-based approaches are commonly used for information extraction and entity recognition, while machine learning algorithms are employed for sentiment analysis, topic modeling, and policy issue identification. The review also highlighted the importance of domain-specific ontologies and labeled training datasets in achieving higher accuracy in policy analysis. Furthermore, we identified research gaps in the application of ontology-driven machine learning techniques for public policy analysis in non-English languages and for policy evaluation. Overall, this SLR provides insights into the current research trends, challenges, and opportunities in ontology-driven machine learning applications for public policy analysis from social media data. A preprint has previously been published by (Abtew et al., 2023).

PDF