AI Glossary / Data engineering and architecture / Intermediate
Generic data model
Definition
A generic data model represents information using broadly reusable kinds of entities and relationships rather than defining a completely separate structure for every business subject. It may describe individuals, classes, properties and relations such as classification or part - whole. Domain-specific concepts can then be expressed through controlled definitions or reference data within that framework.
What it means
A conventional model might create a dedicated structure for each kind of equipment. A more generic model may represent equipment types and their properties as explicit data, allowing new kinds to be described without redesigning every underlying structure. This flexibility moves some modeling work into definitions, constraints and reference data; it does not remove the need for a model. The phrase also appears in product-specific documentation, so the particular modeling approach should be identified when precision matters.
How it works
Designers establish general entity and relationship types and specify how domain vocabulary fits them. For example, a classification relation can connect an individual object to a class, while a part - whole relation connects a component to an assembly. Additional rules constrain which combinations are meaningful. Applications interpret the resulting definitions and facts. A physical implementation may use relational tables, graphs or other storage, but a generic conceptual model is not inherently tied to one storage technology.
Three practical examples
- An engineering information system represents different equipment kinds through shared concepts for objects, classifications and relationships.
- An integration layer uses a common generic representation to exchange entities and attributes from several source systems.
- An AI knowledge resource stores explicit concept definitions and relationships so information from different domains can be mapped into a documented framework.
Think of it this way
Think of a construction kit with reusable connectors and an instruction vocabulary for many structures. Reusing the connectors is convenient, but the instructions still determine what the finished structure means. The analogy is limited because data models must also define identity, constraints and semantics; a flexible connector does not make every possible connection valid.
Worked example
Suppose object P17 is classified as Pump, and object M4 is classified as Motor. A generic relation states that M4 is part of P17. Adding a new class Valve can expand the vocabulary without necessarily adding a new physical table. However, the system still needs to define what Valve means and which properties or relationships are permitted. If someone records “Pump is part of Motor” when they intended an instance-level fact, the model must distinguish classes from individual objects. Generic structure therefore increases the importance of clear typing and validation.
Illustrative case study
Fictional case: Cedar Engineering combines equipment descriptions from several departments. The team uses shared object identifiers and explicit classification relationships, then records domain-specific definitions for pumps and sensors. Reviewers discover that one source uses a model number for an equipment type while another uses a serial number for an individual item. They correct that mapping before loading more records. This fictional case illustrates semantic integration rather than claiming that a generic schema automatically reconciles different meanings.
Common misconceptions and mistakes
Do not equate generic with unstructured, universally correct or free of schema. The structure may be stable while the reference definitions and constraints remain substantial. Do not assume a simple entity - attribute - value table is automatically a complete generic semantic model. Querying and validation can become more complex when domain distinctions are represented as data. Also, flexibility can conceal inconsistent vocabulary unless definitions, identifiers and permitted relationships are governed carefully.
Business use
Businesses may consider generic models when domains evolve or information must be exchanged across diverse systems. Balance adaptability against the effort needed to express constraints, build understandable interfaces and maintain efficient queries. For AI, explicit generic relationships can support knowledge representation, but the usefulness of the result depends on the domain mappings and evidence. Preserve source provenance and model versions so a consumer can understand how a fact was interpreted and whether definitions have changed.
People also ask
Does generic mean the model fits every domain without changes?
No. Domain definitions and constraints still need work, and some requirements may exceed the framework.
Is it necessarily a graph database?
No. The conceptual approach can have different physical implementations.
Can new concepts be added as data?
In many generic approaches, yes, but additions still require clear definitions and validation rules.
Check your understanding
A generic store accepts a relation between any two identifiers. Does that prove the relation is meaningful?
Answer: No. The identifiers' types, the relation's definition and the applicable constraints must make the assertion valid in the intended domain.
Related terms
- Ontology
- Semantic interoperability
- Common data model
- conceptual model
- reference data
- entity–attribute–value model
- classification relation
Learn next
Model one individual object, its class and a component relationship separately. Add a new class and list which definitions and constraints must change even if the storage structure stays the same. Compare the result with a domain-specific table design.
More in Data engineering and architecture
- Fuzzy cognitive map - A fuzzy cognitive map is a model that represents concepts as nodes and graded influences between them as weighted connections.
- Gather/scatter (vector addressing) - In vector addressing, gather and scatter are operations that read or write multiple data elements using a set of memory indices.
- General-purpose computing on graphics processing units - General-purpose computing on graphics processing units, often shortened to GPGPU, uses a GPU to perform computations beyond its traditional graphics role.
- Geo-replication - Geo-replication keeps copies of data in geographically separate locations. It can support regional recovery, distributed access or other resilience goals.
- Georelational data model - A georelational data model represents geographic features by linking spatial information with descriptive attributes.
- Geospatial metadata - Geospatial metadata is information that describes a geographic dataset or resource.
Sources
Primary implementation and research references: Oracle, Explicit Data Set—Generic Data Model, https://docs.oracle.com/html/E54502_01/Content/Technical%20Reference/Generic_Data_Model.htm ; Jordan and colleagues, Re-engineering the ISO 15926 Data Model, https://aise.unisa.edu.au/wp/wp-content/papercite-data/pdf/swese-jordansgms13.pdf . Official product and research excerpts consulted. Generic modelling discussion is distinguished from those specific implementations; full semantic and editorial review remain pending.