In information science an ontology is a type of data model which is used to formally describe a knowledge domain.
There are a number of concepts in this definition which we need to unpack:
- A knowledge domain is an area of interest to researchers in the arts, humanities or social sciences. It corresponds to an aspect of the real world. For example, in the project Beyond the Multiplex, the knowledge domain is film audiences. In the project Intoxicants and Late Modernity, the knowledge domain is the history, culture and economy of intoxicants in England during the early modern period.
- A data model is a description of a knowledge domain which uses entities, categories, taxonomies, rules, and relationships. Importantly, the description has to be capable of being applied to most if not all significant things that occur in the knowledge domain. Significant things are those aspects of the knowledge domain that are important to the researcher.
For example, we can create a data model that describes bookshops: a shop is physical or online; a shop has a name; a shop contains books; every book has a title, author, and a price; a shop has customers who buy books; customers can buy books from more than one bookshop etc.
This data model will broadly apply to all bookshops. It captures the characteristics of bookshops without us having to know which individual bookshop is being described. It does not describe where the customers live because this is not significant for my research, although it might be if somebody else was to define the data model.
Clearly, we all know what a bookshop is, what it contains, and what happens in a bookshop. But a computer does not. If a computer is given a piece of data, such as Waterstones, it does not know that this is a bookshop. If we tell it that Waterstones is the name of a bookshop, it does not know that a bookshop sells books. This is why data models are important. They enable us to structure and categorise data formally and consistently so that computers can understand how different pieces of data are supposed to relate to one another. This is fundamental if we then want the computer to begin telling us things about the data.
A data model is not the same as a database. A data model is an abstract, conceptual representation of a knowledge domain (as we shall see later). A database is a collection of actual data. A database uses a data model, in that the data is stored, organised and labelled according to the rules of the data model. As such, whereas a data model describes a knowledge domain, a database is designed to mirror the data model through the way in which it stores data. Other types of data storage can use data models as well, such as XML files, JSON files and CSV files.
An ontology is a type of data model. There are different types of data models so that a knowledge domain can be described in different ways, such as a graph data model, entity-relationship data model, and document data model. Choosing which type of data model to use is usually determined by what you want to do with the data within a computer system and what types of information you want to get out of it. For example, a graph data model is useful for modelling the relationships within social networks because it is designed for data that is heavily interconnected.
Perhaps the most common type of data model is the relational data model which is used by relational database software, but relational database software can be used to organise, structure and label data using the rules of other types of data models as well, such as an ontology.
An ontology in information science is sometimes referred to as a web ontology, a data ontology or a computational ontology. Broadly speaking, they are all referring to the same idea. An ontology in information science is different to a philosophical ontology, although the former is clearly influenced by the latter and one could see information science ontologies being useful tools for modelling data that might then inform or express the philosophical study of ontology.
URL for this page
https://www.dhi.ac.uk/books/ontology-guide/an-ontology-is-a-data-model