Abstract
Fake news detection on social media has been very challenging, with diverse techniques already implemented based on content of social media data. However, there is a growing need for use of social data context as well for detection techniques. Leveraging semantic technologies capabilities, this research focused on contextual modelling for social media data, with Twitter data utilised as case study. The raw data is aggregated, processed and transformed into a semantic knowledge graph based on RDF data which is subsequently stored within a graph database. With the tweets initially classified as either fake or real using Fakenewsnet application, the knowledge graph facilitates advanced data analytics and potential extension to the social context modelling developed. Furthermore, the modelled data, alongside ensuing inferential data based on class relationships within the knowledge graph constitute a vital input for data analytics with machine learning towards subsequent classification of other news articles as either fake or not.