Development of graph-based knowledge on ransomware attacks using Twitter data

Ransomware is constantly being developed on underground marketplaces, and spreads through Internet, causing damage to individuals’ and businesses’ data. The purpose of this study is to investigate the current issue related to knowledge graphs on ransomware attacks using Twitter data. To Construct a knowledge graph from informal text, three steps need to be followed. Namely, data collection and cleaning, entity extraction, and relation extraction. Although Natural Language Processing techniques are widely used for text representation and modeling, there exist some limitations related to the lack of a dedicated Named Entity recognizer for extracting Ransomware-related entities from unstructured data such as text. Therefore, this article relies on using the ontology approach to construct a ransomware knowledge graph from unstructured data. An improvement to the ontology is done to make it fit the ransomware attack representation based on data captured from the tweets. The Knowledge Graph was developed by extracting relations between entities. In the end, the accuracy of the Knowledge Graph was evaluated using the formal method.