𝐇𝐨𝐰 𝐭𝐨 𝐝𝐞𝐟𝐢𝐧𝐞 𝐜𝐨𝐧𝐜𝐞𝐩𝐭𝐬 𝐟𝐨𝐫 𝐲𝐨𝐮𝐫 𝐎𝐧𝐭𝐨𝐥𝐨𝐠𝐲: 𝐈𝐦𝐚𝐠𝐢𝐧𝐚𝐭𝐢𝐨𝐧 𝐢𝐬 𝐦𝐨𝐫𝐞 𝐢𝐦𝐩𝐨𝐫𝐭𝐚𝐧𝐭 𝐭𝐡𝐚𝐧 𝐤𝐧𝐨𝐰𝐥𝐞𝐝𝐠𝐞 The first step in the Hapsah.org approach for building an ontology is defining concepts and relationships. As an ontologist, you don’t have to posses all the knowledge about the domain you are building an ontology for. You can use the knowledge of domain experts and existing documentation as input. The main thing you need is the ability to deeply imagine what a concept or relationship essentially is. So, as Einstein said: “Imagination is more important than knowledge”. And writing good conceptual definitions requires practice and experience. But of course, there are also some practical guidelines for writing sound conceptual definitions. 𝐂𝐥𝐚𝐫𝐢𝐭𝐲 • Use unambiguous language • Avoid jargon unless it's domain-specific and well-defined • Ensure the definition is understandable to the intended audience. 𝐏𝐫𝐞𝐜𝐢𝐬𝐢𝐨𝐧 • Define the concept narrowly enough to exclude irrelevant instances. • Avoid vague qualifiers like “usually,” “often,” or “some.” 𝐍𝐨𝐧-𝐜𝐢𝐫𝐜𝐮𝐥𝐚𝐫𝐢𝐭𝐲 • Do not define a concept using the term itself or closely related synonyms. 𝐄𝐬𝐬𝐞𝐧𝐭𝐢𝐚𝐥 𝐂𝐡𝐚𝐫𝐚𝐜𝐭𝐞𝐫𝐢𝐬𝐭𝐢𝐜𝐬 • Focus on the necessary and sufficient properties that make the concept what it is. • Avoid listing incidental or contextual attributes. 𝐂𝐨𝐧𝐭𝐞𝐱𝐭𝐮𝐚𝐥 𝐑𝐞𝐥𝐞𝐯𝐚𝐧𝐜𝐞 • Tailor the definition to the domain or use-case (e.g., legal, medical, technical). • Consider cultural or organizational nuances if applicable. 𝐂𝐨𝐧𝐬𝐢𝐬𝐭𝐞𝐧𝐜𝐲 • Ensure alignment with related concepts in the ontology (e.g. hypernyms and hyponyms) • Avoid contradictions or overlaps with other definitions. Follow me to find out how you can create working apps by specifying your ontology (and business rules) within hours. #Ontology #Semantics #Conceptual #Definitions
I completely agree with your remarks about ontologies, but my work on representing and processing semantics in the Knowledge graph environment led me to the idea that ontologies should represent not only "static" semantics, but also the semantics of some processing, also of semantics. That is, ontologies should contain a program code that provides processing in the Knowledge graph environment, or they should be something like a "dynamic construct". I became convinced of this when I was solving the problem of symbolic-neural integration- https://www.researchgate.net/publication/396098608_Time_and_space_in_the_Unified_Knowledge_Graph_environment .
Systemic thinking, common sense, and experience - key success factors for all endeavors
4dNope. There are quite strict rules that guide composition of ontologies. And here we are to remember that an ontology A) always created for particular domain B) is based on that domain taxonomy. I.e. when one builds, let’s say, ontology of cars, there shall be separate definitions for an engine, gear box, transmission and wheels. All these components can not be grouped together ARBITRARILY according to one’s imagination… Otherwise it will be an ontology of one’s cognitive capabilities