Logical Data Modeling - Classification (Taxonomy | Categorization) - Transversal Data Organization
About
Classification is an naming technique for organization where entity or relationship gets classified by giving them a nominal attribute known as a classifier.
The classification of organisms in an ordered system that indicates natural relationships; The science, laws, or principles of classification; systematics; Division into ordered groups or categories.
The output of classification process is an aggregate binary relationship (ie is part of relationship).
In dimensional data, you will find this attributes in dimension.
Articles Related
Design rule
The magical number may be applied to limit:
- the number of member in a class
- the number of class that classify an entity or a association
Number of Class member
According to magical number, the number of value that can be choose for every class attribute should be in the range <math>7 \pm 2</math>
Example of bad class for a department
- Admin Support Center
- Development
- Computer Science
- Data Services
- Digital Media Innovation Program
- Events Center
- Engineering
- Infrastructure
- Enterprise System Integration
- Project Management
If this is the case, you should create a new level in your package (hierarchy)
Number of Class
According to magical number, the number of value that define a association or an entity should be in the range <math>7 \pm 2</math>
For instance, a fact table should not have more than <math>7 \pm 2</math> dimension.
Algorithm
Classification can be performed during data processing operations
- via a rule
- or a classifier algorithm
Documentation / Reference
- Unified Modeling Language (UML) - 2.5.1 - chapter 9