The detection of fake news on social media has become a very active research area. Several approaches and techniques have been proposed and implemented to address the challenge, across diverse technological domains such as NLP (Natural Language Processing) and machine learning. While substantial progress has been made on these, it remains a daunting task due to complexities in its nature. Therefore, it has become pertinent to significantly explore and integrate other technologies to detect fake news on social media. Hence, this research focuses on further exploring and developing native semantic technology solutions for the discourse space. The initial result is a taxonomy classifying socially contextual features for news articles and then Fandet: an OWL ontology for context-based fake news detection by semantically annotating contextual features of news articles and datasets using the ontology. This provides a basis for patterns recognition, analysis, and identification of news articles on social media as either fake or not.