Ontology

🏛️ Ontology

Ontology refers to the essence of things. It originates from the Greek words “onto” (being) and “logos” (study).
Ontology involves understanding entities and their connections and plays a crucial role in Named Entity Recognition (NER) by improving recognition through shared attributes.

🔗 Ontology & Named Entity Recognition (NER)

A mutual attribute between two entities creates an ontology reference, which helps in understanding relationships between them.

🌍 Country Ontology Example

If we take Portugal, France, Spain, Germany, and Italy as entities under the same category (country), they share mutual attributes like:

🏛️ Capital City
💰 Currency
📊 Population
🏙️ Cities
📜 Laws
⚖️ Courts
🏛️ Assembly
💂‍♂️ Army
📈 Economy
🏗️ Architecture
🏰 History
🗣️ Language

Key Insight: If a document mentions architecture more than courts, the context is likely about the architectural history of these countries. This helps NER prioritize relevant entities.

👤 Person Ontology Example

Consider Donald Trump, Joe Biden, Barack Obama, and Hillary Clinton as entities under the same category (person). They share attributes like:

📏 Height
⚖️ Weight
🎂 Birthdate
🏡 Birthplace
🎓 Education
🌎 Nationality
🎩 Political Title
💍 Spouse
📚 Books

Key Insight: These individuals have relationships based on competition and political roles:

Obama vs. Clinton – Democratic Primary Competitors
Clinton vs. Trump (2016) – Presidential Competitors
Biden vs. Trump (2020) – Presidential Competitors
Obama & Biden – Mutual Administration
Clinton supported Biden against Trump

Ontology references:

✅ Being a competitor
✅ Being a supporter
✅ Being a predecessor or successor

These relationships help NER by detecting mutual attributes such as “American Politician”, “U.S. President”, and “Presidential Candidate”.

🧩 Ontology & NLP Technologies

✔️ Relation Detection
✔️ Semantic Role Labeling
✔️ Lexical Semantics
✔️ Named Entity Recognition (NER)