Towards an Automatic Approach to the Design of A Generic Ontology for Information Security
Ines Meriah, Latifa Ben Arfa Rabai, Ridha Khédri · 2021
Ontology modeling plays an important role in broadening the scope of cognitive Artificial Intelligence and Data Science. In this paper, we propose an automatic approach to the design of a generic Information Security ontology. Starting from a comprehensive information security online dictionary, we automatically explore dictionary content to extract security related concepts and their hierarchical relationships. The process is fully automatic and leads to a concept-wide complete ontology. It encompasses web scrapping, concepts classification, pre-processing, archetypes identification, and ontology design. We use classification method and Natural Language Processing techniques to identify and regroup concepts. The obtained ontology is formed by several graph components. Each gives a security concern, which is a concept that is a root in a hierarchical relationship with other concepts. Isolated concepts are called trivial concerns. The ontology contains 2518 concepts and attributes regrouped into 1166 trivial and non-trivial concerns. It is captured in an XML format that can be automatically transformed into a graphML format. These two formats enable the total or partial usage, and the browsing of the ontology. Then, using graph theory terminology, we define the notions of archetypes and concept-attributes. The obtained ontology while it is concept complete, it needs further enrichment with other than hierarchical relationships that can be obtained from security corpuses and information security standards.